{"id":35542,"url":"https://github.com/wangkai930418/awesome-diffusion-categorized","name":"awesome-diffusion-categorized","description":"collection of diffusion model papers categorized by their subareas","projects_count":6714,"last_synced_at":"2026-07-30T03:00:25.067Z","repository":{"id":167301957,"uuid":"642906925","full_name":"wangkai930418/awesome-diffusion-categorized","owner":"wangkai930418","description":"collection of diffusion model papers categorized by their subareas","archived":false,"fork":false,"pushed_at":"2026-03-16T02:13:09.000Z","size":774,"stargazers_count":2216,"open_issues_count":1,"forks_count":102,"subscribers_count":79,"default_branch":"master","last_synced_at":"2026-07-11T03:01:15.905Z","etag":null,"topics":["continual-learning","controlnet","detection","diffusion","diffusion-model","diffusion-models","few-shot","image-edit","inpainting","inversion","segmentation","stable-diffusion","text-guided","tracking"],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/wangkai930418.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2023-05-19T16:06:31.000Z","updated_at":"2026-07-10T04:08:47.000Z","dependencies_parsed_at":"2023-09-22T14:37:02.128Z","dependency_job_id":"6ea9b8c6-7c3e-4975-b959-3b873c3504c1","html_url":"https://github.com/wangkai930418/awesome-diffusion-categorized","commit_stats":{"total_commits":63,"total_committers":1,"mean_commits":63.0,"dds":0.0,"last_synced_commit":"07d05741ec68c6025a539fb06cfd9359ab1e450d"},"previous_names":["wangkai930418/awesome-diffusion-categorized"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/wangkai930418/awesome-diffusion-categorized","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wangkai930418%2Fawesome-diffusion-categorized","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wangkai930418%2Fawesome-diffusion-categorized/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wangkai930418%2Fawesome-diffusion-categorized/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wangkai930418%2Fawesome-diffusion-categorized/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/wangkai930418","download_url":"https://codeload.github.com/wangkai930418/awesome-diffusion-categorized/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/wangkai930418%2Fawesome-diffusion-categorized/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":36057645,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-07-20T02:08:10.276Z","status":"online","status_checked_at":"2026-07-30T02:00:05.956Z","response_time":106,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"created_at":"2024-01-13T12:59:21.199Z","updated_at":"2026-07-30T03:00:25.068Z","primary_language":null,"list_of_lists":false,"displayable":true,"categories":["AR model","VAR model","Super Resolution","Personalized Restoration","Image Restoration","Diffusion Models Inversion","Text Guided Image Editing","Continual Learning","Remove Concept","New Concept Learning","T2I Diffusion Model augmentation","Spatial Control","I2I translation","Segmentation Detection Tracking","Additional conditions","Few-Shot","SD-inpaint","Layout Generation","Text Generation","Accelerate","Video Generation","Video Editing","Video Customization","Image Compression","ID encoder","3D-assets","Try On","Train-Free","Colorization","Face Restoration","Drag Edit","General Concept","Mutiple Concepts","Poster","Decomposition","AR-based"],"sub_categories":["Train-Free"],"readme":"\u003c!-- The superlink doesn't support uppercases --\u003e\n\n\u003c!-- # PostDoc position in LAMP group\n\nWe are looking for postdocs to join [LAMP group](https://groups.google.com/g/ml-news/c/penSOI3751Y?pli=1) working on Diffusion Models. --\u003e\n\n# Awesome Diffusion Categorized\n\n## Contents\n\n- [Visual Illusion](#illusion)\n- [Color](#color-in-generation)\n- [Count](#count-guidance)\n- [Poster](#poster)\n- [Accelerate](#accelerate)\n    - [Train-Free](#train-free)\n    - [AR model](#ar-model)\n    - [VAR model](#var-model)\n- [Image Restoration](#image-restoration)\n    - [Colorization](#colorization)\n    - [Face Restoration](#face-restoration)\n    - [Image Compression](#image-compression)\n    - [Super Resolution](#super-resolution)\n    - [Personalized Restoration](#personalized-restoration)\n- [Storytelling](#storytelling)\n- [Virtual Try On](#try-on)\n- [Drag Edit](#drag-edit)\n- [Text-Guided Editing](#text-guided-image-editing)\n    - [Diffusion Inversion](#diffusion-models-inversion)\n- [Continual Learning](#continual-learning)\n- [Remove Concept](#remove-concept)\n- [In Context Learning](#in-context-learning)\n    - [Multi-Concept](#mutiple-concepts)\n    - [Decompostion](#decomposition)\n    - [ID Encoder](#id-encoder)\n    - [General Personalization](#general-concept)\n    - [AR-based](#ar-based)\n    - [Video Customization](#video-customization)\n\u003c!-- - [T2I augmentation](#t2i-diffusion-model-augmentation)\n    - [Spatial Control](#spatial-control) --\u003e\n- [Image Translation](#i2i-translation)\n- [Seg \u0026 Detect \u0026 Track](#segmentation-detection-tracking)\n- [Adding Conditions](#additional-conditions)\n- [Few-Shot](#few-shot)\n- [Inpainting](#sd-inpaint)\n- [Layout](#layout-generation)\n\u003c!-- - [Text Generation](#text-generation)\n- [Video Generation](#video-generation)\n- [Video Editing](#video-editing) --\u003e\n\n\n# Illusion \n\n**The Art of Deception: Color Visual Illusions and Diffusion Models** \\\n[[CVPR 2025](https://arxiv.org/abs/2412.10122)]\n[[Project](https://openaccess.thecvf.com/content/CVPR2025/html/Gomez-Villa_The_Art_of_Deception_Color_Visual_Illusions_and_Diffusion_Models_CVPR_2025_paper.html)] \n[[Code](https://alviur.github.io/color-illusion-diffusion/)] \n\n**Visual Anagrams: Generating Multi-View Optical Illusions with Diffusion Models** \\\n[[CVPR 2024](https://arxiv.org/abs/2311.17919)]\n[[Project](https://dangeng.github.io/visual_anagrams/)]\n[[Code](https://github.com/dangeng/visual_anagrams)] \n\n**PTDiffusion: Free Lunch for Generating Optical Illusion Hidden Pictures with Phase-Transferred Diffusion Model** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.06186)]\n[[Project](https://xianggao1102.github.io/PTDiffusion_webpage/)]\n[[Code](https://github.com/XiangGao1102/PTDiffusion)] \n\n**Factorized Diffusion: Perceptual Illusions by Noise Decomposition** \\\n[[ECCV 2024](https://arxiv.org/abs/2404.11615)]\n[[Project](https://dangeng.github.io/factorized_diffusion/)]\n[[Code](https://github.com/dangeng/visual_anagrams)] \n\n**Diffusion Illusions: Hiding Images in Plain Sight** \\\n[[Website](https://arxiv.org/abs/2312.03817)]\n[[Project](https://diffusionillusions.com/)]\n[[Code](https://github.com/RyannDaGreat/Diffusion-Illusions)] \n\n**Diffusion-based Visual Anagram as Multi-task Learning** \\\n[[WACV 2025](https://arxiv.org/abs/2412.02693)]\n[[Code](https://github.com/Pixtella/Anagram-MTL)] \n\n**Evaluating Model Perception of Color Illusions in Photorealistic Scenes** \\\n[[Website](https://arxiv.org/abs/2412.06184)]\n[[Code](https://github.com/mao1207/RCID)] \n\n**Illusion3D: 3D Multiview Illusion with 2D Diffusion Priors** \\\n[[Website](https://arxiv.org/abs/2412.09625)]\n[[Project](https://3d-multiview-illusion.github.io/)] \n\n\n# Color in Generation\n\n**ColorPeel: Color Prompt Learning with Diffusion Models via Color and Shape Disentanglement** \\\n[[ECCV 2024](https://arxiv.org/abs/2407.07197)] \n[[Project](https://moatifbutt.github.io/colorpeel/)]\n[[Code](https://github.com/moatifbutt/color-peel)]\n\n**Free-Lunch Color-Texture Disentanglement for Stylized Image Generation** \\\n[[NeurIPS 2025](https://arxiv.org/abs/2503.14275v1)] \n[[Project](https://deepffff.github.io/sadis.github.io/)] \n[[Code](https://github.com/deepffff/SADis)]\n\n**Not Every Gift Comes in Gold Paper or with a Red Ribbon: Exploring Color Perception in Text-to-Image Models** \\\n[[WACV 2026](https://arxiv.org/abs/2508.19791)]\n[[Project](https://tau-vailab.github.io/color-edit/)] \n[[Code](https://github.com/TAU-VAILab/color-edit)]\n\n**Color Conditional Generation with Sliced Wasserstein Guidance** \\\n[[NeurIPS 2025](https://arxiv.org/abs/2503.19034)]\n[[Code](https://github.com/alobashev/sw-guidance/)]\n\n**Leveraging Semantic Attribute Binding for Free-Lunch Color Control in Diffusion Models** \\\n[[WACV 2026](https://arxiv.org/abs/2503.09864)]\n[[Project](https://hecoding.github.io/colorwave-page/)]\n\n**Evaluating Model Perception of Color Illusions in Photorealistic Scenes** \\\n[[Website](https://arxiv.org/abs/2412.06184)]\n[[Code](https://github.com/mao1207/RCID)]\n\n\n\n**Exploring Palette based Color Guidance in Diffusion Models** \\\n[[ACM MM 2025](https://arxiv.org/abs/2508.08754)]\n\n**Color Me Correctly: Bridging Perceptual Color Spaces and Text Embeddings for Improved Diffusion Generation** \\\n[[ACM MM 2025](https://arxiv.org/abs/2509.10058)]\n\n**GenColorBench: A Color Evaluation Benchmark for Text-to-Image Generation Models** \\\n[[Website](https://arxiv.org/abs/2510.20586)]\n\n**Training-Free Text-Guided Color Editing with Multi-Modal Diffusion Transformer** \\\n[[Website](https://arxiv.org/abs/2508.09131)]\n\n**DiffBrush:Just Painting the Art by Your Hands** \\\n[[Website](https://arxiv.org/abs/2502.20904)]\n\n**GenColor: Generative Color-Concept Association in Visual Design** \\\n[[Website](https://arxiv.org/abs/2503.03236)]\n\n**Training-free Color-Style Disentanglement for Constrained Text-to-Image Synthesis** \\\n[[Website](https://arxiv.org/abs/2409.02429)]\n\n**Color encoding in Latent Space of Stable Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2512.09477)]\n\n# Count Guidance\n\n**Make It Count: Text-to-Image Generation with an Accurate Number of Objects** \\\n[[CVPR 2025](https://arxiv.org/abs/2406.10210)] \n[[Project](https://make-it-count-paper.github.io//)] \n[[Code](https://github.com/Litalby1/make-it-count)]\n\n**Detection-Driven Object Count Optimization for Text-to-Image Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2408.11721v2)] \n[[Project](https://ozzafar.github.io/count_token/)] \n[[Code](https://github.com/ozzafar/discriminative_class_tokens_for_counting)]\n\n**CountCluster: Training-Free Object Quantity Guidance with Cross-Attention Map Clustering for Text-to-Image Generation** \\\n[[Website](https://arxiv.org/abs/2508.10710)] \n[[Code](https://github.com/JoohyeonL22/CountCluster)] \n\n**YOLO-Count: Differentiable Object Counting for Text-to-Image Generation** \\\n[[ICCV 2025](https://arxiv.org/abs/2508.00728)] \n\n**CountDiffusion: Text-to-Image Synthesis with Training-Free Counting-Guidance Diffusion** \\\n[[Website](https://arxiv.org/abs/2505.04347)] \n\n\n# Poster\n\n**PosterCraft: Rethinking High-Quality Aesthetic Poster Generation in a Unified Framework** \\\n[[Website](https://arxiv.org/abs/2506.10741)] \n[[Project](https://ephemeral182.github.io/PosterCraft/)] \n[[Code](https://github.com/Ephemeral182/PosterCraft)]\n\n**CreatiPoster: Towards Editable and Controllable Multi-Layer Graphic Design Generation** \\\n[[Website](https://arxiv.org/abs/2506.10890)] \n[[Code](https://github.com/graphic-design-ai/creatiposter)] \n\n**PosterMaker: Towards High-Quality Product Poster Generation with Accurate Text Rendering** \\\n[[CVPR 2025](https://arxiv.org/abs/2504.06632)]\n[[Project](https://poster-maker.github.io/)] \n\n**POSTA: A Go-to Framework for Customized Artistic Poster Generation** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.14908)] \n[[Project](https://haoyuchen.com/POSTA)] \n\n**DreamPoster: A Unified Framework for Image-Conditioned Generative Poster Design** \\\n[[Website](https://arxiv.org/abs/2507.04218)] \n[[Project](https://dreamposter.github.io/)] \n\n**LogoDiffuser: Training-Free Multilingual Logo Generation and Stylization via Letter-Aware Attention Control** \\\n[[Website](https://arxiv.org/abs/2603.09759)] \n\n# Accelerate\n\n**PIXART-δ: Fast and Controllable Image Generation with Latent Consistency Models** \\\n[[ICLR 2024 Spotlight](https://arxiv.org/abs/2401.05252)]\n[[Diffusers 1](https://huggingface.co/docs/diffusers/main/en/api/pipelines/pixart)]\n[[Diffusers 2](https://huggingface.co/PixArt-alpha/PixArt-XL-2-1024-MS)]\n[[Project](https://pixart-alpha.github.io/)]\n[[Code](https://github.com/PixArt-alpha/PixArt-alpha)]\n\n**SDXL-Turbo: Adversarial Diffusion Distillation** \\\n[[Website](https://arxiv.org/abs/2311.17042)]\n[[Diffusers 1](https://huggingface.co/stabilityai/sdxl-turbo)]\n[[Diffusers 2](https://huggingface.co/docs/diffusers/en/using-diffusers/sdxl_turbo)]\n[[Project](https://huggingface.co/stabilityai)]\n[[Code](https://github.com/Stability-AI/generative-models)]\n\n**Trajectory Consistency Distillation: Improved Latent Consistency Distillation by Semi-Linear Consistency Function with Trajectory Mapping** \\\n[[Website](https://arxiv.org/abs/2405.14867)]\n[[Diffusers 1](https://huggingface.co/h1t/TCD-SDXL-LoRA)]\n[[Diffusers 2](https://huggingface.co/docs/diffusers/en/using-diffusers/inference_with_tcd_lora)]\n[[Project](https://tianweiy.github.io/dmd2/)]\n[[Code](https://github.com/jabir-zheng/TCD)]\n\n**LCM-LoRA: A Universal Stable-Diffusion Acceleration Module** \\\n[[Website](https://arxiv.org/abs/2311.05556)]\n[[Diffusers](https://huggingface.co/docs/diffusers/en/using-diffusers/inference_with_lcm?lcm-lora=LCM-LoRA#lora)]\n[[Project](https://latent-consistency-models.github.io/)]\n[[Code](https://github.com/luosiallen/latent-consistency-model)]\n\n**Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference** \\\n[[Website](https://arxiv.org/abs/2310.04378)]\n[[Project](https://huggingface.co/docs/diffusers/api/pipelines/latent_consistency_models)]\n[[Code](https://github.com/luosiallen/latent-consistency-model)]\n\n**DMD2: Improved Distribution Matching Distillation for Fast Image Synthesis** \\\n[[NeurIPS 2024 Oral](https://arxiv.org/abs/2405.14867)]\n[[Project](https://tianweiy.github.io/dmd2/)]\n[[Code](https://github.com/tianweiy/DMD2)]\n\n**DMD1: One-step Diffusion with Distribution Matching Distillation** \\\n[[CVPR 2024](https://arxiv.org/abs/2311.18828)]\n[[Project](https://tianweiy.github.io/dmd/)]\n[[Code](https://github.com/devrimcavusoglu/dmd)]\n\n**Tortoise and Hare Guidance: Accelerating Diffusion Model Inference with Multirate Integration** \\\n[[NeurIPS 2025](https://arxiv.org/abs/2511.04117)]\n[[Project](https://yhlee-add.github.io/THG/)]\n[[Code](https://github.com/yhlee-add/THG)]\n\n**Consistency Models** \\\n[[ICML 2023](https://proceedings.mlr.press/v202/song23a.html)]\n[[Diffusers](https://huggingface.co/docs/diffusers/main/en/api/pipelines/consistency_models)]\n[[Code](https://github.com/openai/consistency_models)]\n\n**SwiftBrush: One-Step Text-to-Image Diffusion Model with Variational Score Distillation** \\\n[[CVPR 2024](https://arxiv.org/abs/2312.05239)]\n[[Project](https://vinairesearch.github.io/SwiftBrush/)]\n[[Code](https://github.com/VinAIResearch/SwiftBrush)]\n\n**SwiftBrush V2: Make Your One-Step Diffusion Model Better Than Its Teacher** \\\n[[ECCV 2024](https://arxiv.org/abs/2408.14176)]\n[[Project](https://swiftbrushv2.github.io/)]\n[[Code](https://github.com/VinAIResearch/SwiftBrush)]\n\n**CoDi: Conditional Diffusion Distillation for Higher-Fidelity and Faster Image Generation** \\\n[[CVPR 2024](https://arxiv.org/abs/2310.01407)]\n[[Project](https://fast-codi.github.io/)]\n[[Code](https://github.com/fast-codi/CoDi)]\n\n**PCM : Phased Consistency Model** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2405.18407)]\n[[Project](https://g-u-n.github.io/projects/pcm/)]\n[[Code](https://github.com/G-U-N/Phased-Consistency-Model)]\n\n**Motion Consistency Model: Accelerating Video Diffusion with Disentangled Motion-Appearance Distillation** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2406.06890)]\n[[Project](https://yhzhai.github.io/mcm/)]\n[[Code](https://github.com/yhZhai/mcm)]\n\n**KOALA: Empirical Lessons Toward Memory-Efficient and Fast Diffusion Models for Text-to-Image Synthesis** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2312.04005)]\n[[Project](https://youngwanlee.github.io/KOALA/)]\n[[Code](https://github.com/youngwanLEE/sdxl-koala)]\n\n**Random Conditioning with Distillation for Data-Efficient Diffusion Model Compression** \\\n[[CVPR 2025](https://arxiv.org/abs/2504.02011)]\n[[Project](https://dohyun-as.github.io/Random-Conditioning/)]\n[[Code](https://github.com/dohyun-as/Random-Conditioning)]\n\n**DIMO:Distilling Masked Diffusion Models into One-step Generator** \\\n[[Website](https://arxiv.org/abs/2503.15457)]\n[[Project](https://yuanzhi-zhu.github.io/DiMO/)]\n[[Code](https://github.com/yuanzhi-zhu/DiMO)]\n\n**Flash Diffusion: Accelerating Any Conditional Diffusion Model for Few Steps Image Generation** \\\n[[Website](https://arxiv.org/abs/2406.02347)]\n[[Project](https://gojasper.github.io/flash-diffusion-project/)]\n[[Code](https://github.com/gojasper/flash-diffusion)]\n\n**Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2411.19108)]\n[[Project](https://liewfeng.github.io/TeaCache/)]\n[[Code](https://github.com/LiewFeng/TeaCache)]\n\n**You Only Sample Once: Taming One-Step Text-to-Image Synthesis by Self-Cooperative Diffusion GANs** \\\n[[Website](https://arxiv.org/abs/2403.12931)]\n[[Project](https://yoso-t2i.github.io/)]\n[[Code](https://github.com/Luo-Yihong/YOSO)]\n\n**PeRFlow: Piecewise Rectified Flow as Universal Plug-and-Play Accelerator** \\\n[[Website](https://arxiv.org/abs/2405.07510)]\n[[Project](https://piecewise-rectified-flow.github.io/)]\n[[Code](https://github.com/magic-research/piecewise-rectified-flow)]\n\n**Scale-wise Distillation of Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2503.16397)]\n[[Project](https://yandex-research.github.io/swd/)]\n[[Code](https://github.com/yandex-research/swd)]\n\n**Simplifying, Stabilizing and Scaling Continuous-Time Consistency Models** \\\n[[Website](https://doi.org/10.48550/arXiv.2410.11081)]\n[[Project](https://openai.com/index/simplifying-stabilizing-and-scaling-continuous-time-consistency-models/)]\n[[Code](https://github.com/xandergos/sCM-mnist)]\n\n**Adaptive Caching for Faster Video Generation with Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2411.02397)]\n[[Project](https://adacache-dit.github.io/)]\n[[Code](https://github.com/AdaCache-DiT/AdaCache)]\n\n**FasterCache: Training-Free Video Diffusion Model Acceleration with High Quality** \\\n[[Website](https://arxiv.org/abs/2410.19355)]\n[[Project](https://vchitect.github.io/FasterCache/)]\n[[Code](https://github.com/Vchitect/FasterCache)]\n\n**Learning Few-Step Diffusion Models by Trajectory Distribution Matching** \\\n[[Website](https://arxiv.org/abs/2503.06674)]\n[[Project](https://tdm-t2x.github.io/)]\n[[Code](https://github.com/Luo-Yihong/TDM)]\n\n**SDXS: Real-Time One-Step Latent Diffusion Models with Image Conditions** \\\n[[Website](https://arxiv.org/abs/2403.16627)]\n[[Project](https://idkiro.github.io/sdxs/)]\n[[Code](https://github.com/IDKiro/sdxs)]\n\n**Reward Guided Latent Consistency Distillation** \\\n[[Website](https://arxiv.org/abs/2403.11027)]\n[[Project](https://rg-lcd.github.io/)]\n[[Code](https://github.com/Ji4chenLi/rg-lcd)]\n\n**T-Stitch: Accelerating Sampling in Pre-Trained Diffusion Models with Trajectory Stitching** \\\n[[Website](https://arxiv.org/abs/2402.14167)]\n[[Project](https://t-stitch.github.io/)]\n[[Code](https://github.com/NVlabs/T-Stitch)]\n\n**Accelerating Diffusion Sampling via Exploiting Local Transition Coherence** \\\n[[Website](https://arxiv.org/abs/2503.09675)]\n[[Project](https://zhushangwen.github.io/LTC-accel.io/)]\n[[Code](https://github.com/zhushangwen/LTC-Accel)]\n\n**AccVideo: Accelerating Video Diffusion Model with Synthetic Dataset** \\\n[[Website](https://arxiv.org/abs/2503.19462)]\n[[Project](https://aejion.github.io/accvideo/)]\n[[Code](https://github.com/aejion/AccVideo/)]\n\n**One-Step Offline Distillation of Diffusion-based Models via Koopman Modeling** \\\n[[Website](https://arxiv.org/abs/2505.13358)]\n[[Project](https://sites.google.com/view/koopman-distillation-model/)]\n[[Code](https://github.com/azencot-group/KDM)]\n\n**Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding** \\\n[[Website](https://arxiv.org/abs/2505.22618)]\n[[Project](https://nvlabs.github.io/Fast-dLLM/)]\n[[Code](https://github.com/NVlabs/Fast-dLLM)]\n\n**MagCache: Fast Video Generation with Magnitude-Aware Cache** \\\n[[Website](https://arxiv.org/abs/2506.09045)]\n[[Project](https://zehong-ma.github.io/MagCache/)]\n[[Code](https://github.com/Zehong-Ma/MagCache)]\n\n**Evolutionary Caching to Accelerate Your Off-the-Shelf Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2506.15682)]\n[[Project](https://aniaggarwal.github.io/ecad/)]\n[[Code](https://github.com/aniaggarwal/ecad)]\n\n**Less is Enough: Training-Free Video Diffusion Acceleration via Runtime-Adaptive Caching** \\\n[[Website](https://arxiv.org/abs/2507.02860)]\n[[Project](https://h-embodvis.github.io/EasyCache/)]\n[[Code](https://github.com/H-EmbodVis/EasyCache)]\n\n**Distilling Diversity and Control in Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2503.10637)]\n[[Project](https://distillation.baulab.info/)]\n[[Code](https://github.com/rohitgandikota/distillation)]\n\n**SANA-Sprint: One-Step Diffusion with Continuous-Time Consistency Distillation** \\\n[[Website](https://arxiv.org/abs/2503.09641)]\n[[Project](https://nvlabs.github.io/Sana/Sprint/)]\n[[Code](https://github.com/NVlabs/Sana)]\n\n**LeMiCa: Lexicographic Minimax Path Caching for Efficient Diffusion-Based Video Generation** \\\n[[Website](https://arxiv.org/abs/2511.00090)]\n[[Project](https://unicomai.github.io/LeMiCa/)]\n[[Code](https://github.com/UnicomAI/LeMiCa)]\n\n**Glance: Accelerating Diffusion Models with 1 Sample** \\\n[[Website](https://arxiv.org/abs/2512.02899)]\n[[Project](https://zhuobaidong.github.io/Glance/)]\n[[Code](https://github.com/zhuobaidong/Glance)]\n\n**No Cache Left Idle: Accelerating diffusion model via Extreme-slimming Caching** \\\n[[Website](https://arxiv.org/abs/2512.12604)]\n[[Project](https://thu-accdiff.github.io/xslim-page/)]\n[[Code](https://github.com/THU-AccDiff/xslim/]\n\n**Relational Feature Caching for Accelerating Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2602.19506)]\n[[Project](https://cvlab.yonsei.ac.kr/projects/RFC/)]\n[[Code](https://github.com/cvlab-yonsei]\n\n**One-Way Ticket:Time-Independent Unified Encoder for Distilling Text-to-Image Diffusion Models** \\\n[[CVPR 2025](https://arxiv.org/abs/2505.21960)]\n\u003c!-- [[Project](https://openaccess.thecvf.com/content/CVPR2025/html/Li_One-Way_Ticket_Time-Independent_Unified_Encoder_for_Distilling_Text-to-Image_Diffusion_Models_CVPR_2025_paper.html)] --\u003e\n[[Code](https://github.com/sen-mao/Loopfree)]\n\n**Relational Diffusion Distillation for Efficient Image Generation** \\\n[[ACM MM 2024 (Oral)](https://arxiv.org/abs/2410.07679)]\n[[Code](https://github.com/cantbebetter2/RDD)]\n\n**Autoregressive Distillation of Diffusion Transformers** \\\n[[CVPR 2025 Oral](https://arxiv.org/abs/2504.11295)]\n[[Code](https://github.com/alsdudrla10/ARD)]\n\n**UFOGen: You Forward Once Large Scale Text-to-Image Generation via Diffusion GANs** \\\n[[CVPR 2024](https://arxiv.org/abs/2311.09257)]\n[[Code](https://github.com/xuyanwu/SIDDMs-UFOGen)]\n\n**Accelerating Diffusion Transformer via Increment-Calibrated Caching with Channel-Aware Singular Value Decomposition** \\\n[[CVPR 2025](https://arxiv.org/abs/2505.05829)]\n[[Code](https://github.com/ccccczzy/icc)]\n\n**ToM: Decider-Guided Dynamic Token Merging for Accelerating Diffusion MLLMs** \\\n[[AAAI 2026](https://arxiv.org/abs/2511.12280)]\n[[Code](https://github.com/bcmi/D3ToM-Diffusion-MLLM)]\n\n**SADA: Stability-guided Adaptive Diffusion Acceleration** \\\n[[ICML 2025](https://arxiv.org/abs/2507.17135)]\n[[Code](https://github.com/Ting-Justin-Jiang/sada-icml)]\n\n**SlimFlow: Training Smaller One-Step Diffusion Models with Rectified Flow** \\\n[[ECCV 2024](https://arxiv.org/abs/2407.12718)]\n[[Code](https://github.com/yuanzhi-zhu/SlimFlow)]\n\n**Accelerating Image Generation with Sub-path Linear Approximation Model** \\\n[[ECCV 2024](https://arxiv.org/abs/2404.13903)]\n[[Code](https://github.com/MCG-NJU/SPLAM)]\n\n**Diff-Instruct: A Universal Approach for Transferring Knowledge From Pre-trained Diffusion Models** \\\n[[NeurIPS 2023](https://arxiv.org/abs/2305.18455)]\n[[Code](https://github.com/pkulwj1994/diff_instruct)]\n\n**Fast and Memory-Efficient Video Diffusion Using Streamlined Inference** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2411.01171)]\n[[Code](https://github.com/wuyushuwys/FMEDiffusion)]\n\n**Accelerating Diffusion via Hybrid Data-Pipeline Parallelism Based on Conditional Guidance Scheduling** \\\n[[CVPR 2026](https://arxiv.org/abs/2602.21760)]\n[[Code](https://github.com/kaist-dmlab/Hybridiff)]\n\n**A Simple Early Exiting Framework for Accelerated Sampling in Diffusion Models** \\\n[[ICML 2024](https://arxiv.org/abs/2408.05927)]\n[[Code](https://github.com/taehong-moon/ee-diffusion)]\n\n**Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generation** \\\n[[ICML 2024](https://arxiv.org/abs/2404.04057)]\n[[Code](https://github.com/mingyuanzhou/SiD)]\n\n**On the Trajectory Regularity of ODE-based Diffusion Sampling** \\\n[[ICML 2024](https://arxiv.org/abs/2405.11326)]\n[[Code](https://github.com/zju-pi/diff-sampler)]\n\n**InstaFlow: One Step is Enough for High-Quality Diffusion-Based Text-to-Image Generation** \\\n[[ICLR 2024](https://arxiv.org/abs/2309.06380)]\n[[Code](https://github.com/gnobitab/instaflow)]\n\n**Improved Training Technique for Latent Consistency Models** \\\n[[ICLR 2025](https://arxiv.org/abs/2502.01441)]\n[[Code](https://github.com/quandao10/sLCT/)]\n\n**ProCache: Constraint-Aware Feature Caching with Selective Computation for Diffusion Transformer Acceleration** \\\n[[AAAI 2026](https://arxiv.org/abs//2512.17298)]\n[[Code](https://github.com/macovaseas/ProCache)]\n\n**Compute Only 16 Tokens in One Timestep: Accelerating Diffusion Transformers with Cluster-Driven Feature Caching** \\\n[[ACM MM 2025](https://arxiv.org/abs/2509.10312)]\n[[Code](https://github.com/Shenyi-Z/Cache4Diffusion)]\n\n**CacheQuant: Comprehensively Accelerated Diffusion Models** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.01323)]\n[[Code](https://github.com/BienLuky/CacheQuant)]\n\n**SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion Models** \\\n[[CVPR 2026](https://arxiv.org/abs/2602.18993)]\n[[Code](https://github.com/jiwoogit/SeaCache)]\n\n**Accelerating Vision Diffusion Transformers with Skip Branches** \\\n[[Website](https://arxiv.org/abs/2411.17616)]\n[[Code](https://github.com/OpenSparseLLMs/Skip-DiT)]\n\n**Accelerating Diffusion Transformers with Dual Feature Caching** \\\n[[Website](https://arxiv.org/abs/2412.18911)]\n[[Code](https://github.com/shenyi-z/duca)]\n\n**From Reusing to Forecasting: Accelerating Diffusion Models with TaylorSeers** \\\n[[Website](https://arxiv.org/abs/2503.06923)]\n[[Code](https://github.com/Shenyi-Z/TaylorSeer)]\n\n**Exposure Bias Reduction for Enhancing Diffusion Transformer Feature Caching** \\\n[[Website](https://arxiv.org/abs/2503.07120)]\n[[Code](https://github.com/aSleepyTree/EB-Cache)]\n\n**One Step Diffusion via Shortcut Models** \\\n[[Website](https://arxiv.org/abs/2410.12557)]\n[[Code](https://github.com/kvfrans/shortcut-models)]\n\n**DuoDiff: Accelerating Diffusion Models with a Dual-Backbone Approach** \\\n[[Website](https://arxiv.org/abs/2410.09633)]\n[[Code](https://github.com/razvanmatisan/duodiff)]\n\n**DraftAttention: Fast Video Diffusion via Low-Resolution Attention Guidance** \\\n[[Website](https://arxiv.org/abs/2505.14708)]\n[[Code](https://github.com/shawnricecake/draft-attention)]\n\n**A Closer Look at Time Steps is Worthy of Triple Speed-Up for Diffusion Model Training** \\\n[[Website](https://arxiv.org/abs/2405.17403)]\n[[Code](https://github.com/nus-hpc-ai-lab/speed)]\n\n**Stable Consistency Tuning: Understanding and Improving Consistency Models** \\\n[[Website](https://arxiv.org/abs/2410.18958)]\n[[Code](https://github.com/G-U-N/Stable-Consistency-Tuning)]\n\n**SpeedUpNet: A Plug-and-Play Adapter Network for Accelerating Text-to-Image Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2312.08887)]\n[[Code](https://github.com/williechai/speedup-plugin-for-stable-diffusions)]\n\n**Learning-to-Cache: Accelerating Diffusion Transformer via Layer Caching** \\\n[[Website](https://arxiv.org/abs/2406.01733)]\n[[Code](https://github.com/horseee/learning-to-cache)]\n\n**SDXL-Lightning: Progressive Adversarial Diffusion Distillation** \\\n[[Website](https://arxiv.org/abs/2402.13929)]\n[[Code](https://huggingface.co/ByteDance/SDXL-Lightning)]\n\n**Distribution Backtracking Builds A Faster Convergence Trajectory for Diffusion Distillation** \\\n[[Website](https://arxiv.org/abs/2408.15991)]\n[[Code](https://github.com/SYZhang0805/DisBack)]\n\n**Long and Short Guidance in Score identity Distillation for One-Step Text-to-Image Generation** \\\n[[Website](https://arxiv.org/abs/2406.01561)]\n[[Code](https://github.com/mingyuanzhou/SiD-LSG)]\n\n**Diffusion Models Are Innate One-Step Generators** \\\n[[Website](https://arxiv.org/abs/2405.20750)]\n[[Code](https://github.com/Zyriix/GDD)]\n\n**Optimal Stepsize for Diffusion Sampling** \\\n[[Website](https://arxiv.org/abs/2503.21774)]\n[[Code](https://github.com/bebebe666/OptimalSteps)]\n\n**Model Reveals What to Cache: Profiling-Based Feature Reuse for Video Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2504.03140)]\n[[Code](https://github.com/GeekGuru123/ProfilingDiT/tree/main)]\n\n**Few-Step Diffusion via Score identity Distillation** \\\n[[Website](https://arxiv.org/abs/2505.12674)]\n[[Code](https://github.com/mingyuanzhou/SiD-LSG)]\n\n**FastCache: Fast Caching for Diffusion Transformer Through Learnable Linear Approximation** \\\n[[Website](https://arxiv.org/abs/2505.20353)]\n[[Code](https://github.com/NoakLiu/FastCache-xDiT)]\n\n**SenseFlow: Scaling Distribution Matching for Flow-based Text-to-Image Distillation** \\\n[[Website](https://arxiv.org/abs/2506.00523)]\n[[Code](https://github.com/XingtongGe/SenseFlow)]\n\n**Sparse-vDiT: Unleashing the Power of Sparse Attention to Accelerate Video Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2506.03065)]\n[[Code](https://github.com/Peyton-Chen/Sparse-vDiT)]\n\n**Morse: Dual-Sampling for Lossless Acceleration of Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2506.18251)]\n[[Code](https://github.com/deep-optimization/Morse)]\n\n**SpeCa: Accelerating Diffusion Transformers with Speculative Feature Caching** \\\n[[Website](https://arxiv.org/abs/2509.11628)]\n[[Code](https://github.com/Shenyi-Z/Cache4Diffusion/)]\n\n**QuantSparse: Comprehensively Compressing Video Diffusion Transformer with Model Quantization and Attention Sparsification** \\\n[[Website](https://arxiv.org/abs/2509.23681)]\n[[Code](https://github.com/wlfeng0509/QuantSparse)]\n\n**DC-Gen: Post-Training Diffusion Acceleration with Deeply Compressed Latent Space** \\\n[[Website](https://arxiv.org/abs/2509.25180)]\n[[Code](https://github.com/dc-ai-projects/DC-Gen)]\n\n**QuantSparse: Comprehensively Compressing Video Diffusion Transformer with Model Quantization and Attention Sparsification** \\\n[[Website](https://arxiv.org/abs/2509.23681)]\n[[Code](https://github.com/wlfeng0509/QuantSparse)]\n\n**Towards Better \u0026 Faster Autoregressive Image Generation: From the Perspective of Entropy** \\\n[[Website](https://arxiv.org/abs/2510.09012)]\n[[Code](https://github.com/krennic999/ARsample)]\n\n**pi-Flow: Policy-Based Few-Step Generation via Imitation Distillation** \\\n[[Website](https://arxiv.org/abs/2510.14974)]\n[[Code](https://github.com/Lakonik/piFlow)]\n\n**Towards One-step Causal Video Generation via Adversarial Self-Distillation** \\\n[[Website](https://arxiv.org/abs/2511.01419)]\n[[Code](https://github.com/BigAandSmallq/SAD)]\n\n**RedVTP: Training-Free Acceleration of Diffusion Vision-Language Models Inference via Masked Token-Guided Visual Token Pruning** \\\n[[Website](https://arxiv.org/abs/2511.12428)]\n[[Code](https://github.com/Blacktower27/RedVTP)]\n\n**Decoupled DMD: CFG Augmentation as the Spear, Distribution Matching as the Shield** \\\n[[Website](https://arxiv.org/abs/2511.22677)]\n[[Code](https://github.com/Tongyi-MAI/Z-Image)]\n\n**TurboDiffusion: Accelerating Video Diffusion Models by 100-200 Times** \\\n[[Website](https://arxiv.org/abs/2512.16093)]\n[[Code](https://github.com/thu-ml/TurboDiffusion)]\n\n**CorGi: Contribution-Guided Block-Wise Interval Caching for Training-Free Acceleration of Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2512.24195)]\n[[Code](https://github.com/CASL-KU/CorGi)]\n\n**ArcFlow: Unleashing 2-Step Text-to-Image Generation via High-Precision Non-Linear Flow Distillation** \\\n[[Website](https://arxiv.org/abs/2602.09014)]\n[[Code](https://github.com/pnotp/ArcFlow)]\n\n**Jano: Adaptive Diffusion Generation with Early-stage Convergence Awareness** \\\n[[Website](https://arxiv.org/abs/2603.00519)]\n[[Code](https://github.com/chen-yy20/Jano)]\n\n**SODA: Sensitivity-Oriented Dynamic Acceleration for Diffusion Transformer** \\\n[[Website](https://arxiv.org/abs/2603.07057)]\n[[Code](https://github.com/leaves162/SODA)]\n\n**TDM-R1: Reinforcing Few-Step Diffusion Models with Non-Differentiable Reward** \\\n[[Website](https://arxiv.org/abs/2603.07700)]\n[[Code](https://github.com/Luo-Yihong/TDM-R1)]\n\n**Distilling Diffusion Models into Conditional GANs** \\\n[[ECCV 2024](https://arxiv.org/abs/2405.05967)]\n[[Project](https://mingukkang.github.io/Diffusion2GAN/)]\n\n**Shortcutting Pre-trained Flow Matching Diffusion Models is Almost Free Lunch** \\\n[[NeurIPS 2025](https://arxiv.org/abs/2510.17858)]\n[[Project](https://shortcutfm.github.io/)]\n\n**Cache Me if You Can: Accelerating Diffusion Models through Block Caching** \\\n[[CVPR 2024](https://arxiv.org/abs/2312.03209)]\n[[Project](https://fwmb.github.io/blockcaching/)]\n\n**Plug-and-Play Diffusion Distillation** \\\n[[CVPR 2024](https://arxiv.org/abs/2406.01954)]\n[[Project](https://5410tiffany.github.io/plug-and-play-diffusion-distillation.github.io/)]\n\n**SnapFusion: Text-to-Image Diffusion Model on Mobile Devices within Two Seconds** \\\n[[NeurIPS 2023](https://arxiv.org/abs/2306.00980)]\n[[Project](https://snap-research.github.io/SnapFusion/)]\n\n**One-step Diffusion Models with f-Divergence Distribution Matching** \\\n[[Website](https://arxiv.org/abs/2502.15681)]\n[[Project](https://research.nvidia.com/labs/genair/f-distill/)]\n\n**MagicDistillation: Weak-to-Strong Video Distillation for Large-Scale Portrait Few-Step Synthesis** \\\n[[Website](https://arxiv.org/abs/2503.13319)]\n[[Project](https://w2svd.github.io/W2SVD/)]\n\n**Diffusion Adversarial Post-Training for One-Step Video Generation** \\\n[[Website](https://arxiv.org/abs/2501.08316)]\n[[Project](https://seaweed-apt.com/)]\n\n**SNOOPI: Supercharged One-step Diffusion Distillation with Proper Guidance** \\\n[[Website](https://arxiv.org/abs/2412.02687)]\n[[Project](https://snoopi-onestep.github.io/)]\n\n**NitroFusion: High-Fidelity Single-Step Diffusion through Dynamic Adversarial Training** \\\n[[Website](https://arxiv.org/abs/2412.02030)]\n[[Project](https://chendaryen.github.io/NitroFusion.github.io/)]\n\n**Truncated Consistency Models** \\\n[[Website](https://arxiv.org/abs/2410.14895)]\n[[Project](https://truncated-cm.github.io/)]\n\n**Multi-student Diffusion Distillation for Better One-step Generators** \\\n[[Website](https://arxiv.org/abs/2410.23274)]\n[[Project](https://research.nvidia.com/labs/toronto-ai/MSD/index_hidden.html)]\n\n**Effortless Efficiency: Low-Cost Pruning of Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2412.02852)]\n[[Project](https://yangzhang-v5.github.io/EcoDiff/)]\n\n**SnapGen: Taming High-Resolution Text-to-Image Models for Mobile Devices with Efficient Architectures and Training** \\\n[[Website](https://arxiv.org/abs/2412.09619)]\n[[Project](https://snap-research.github.io/snapgen/)]\n\n**SnapGen-V: Generating a Five-Second Video within Five Seconds on a Mobile Device** \\\n[[Website](https://arxiv.org/abs/2412.10494)]\n[[Project](https://snap-research.github.io/snapgen-v/)]\n\n**Align Your Flow: Scaling Continuous-Time Flow Map Distillation** \\\n[[Website](https://arxiv.org/abs/2506.14603)]\n[[Project](https://research.nvidia.com/labs/toronto-ai/AlignYourFlow/)]\n\n**Training-Free Motion Customization for Distilled Video Generators with Adaptive Test-Time Distillation** \\\n[[Website](https://arxiv.org/abs/2506.19348)]\n[[Project](https://euminds.github.io/motionecho/)]\n\n**Forecasting When to Forecast: Accelerating Diffusion Models with Confidence-Gated Taylor** \\\n[[Website](https://arxiv.org/abs/2508.02240)]\n[[Project](https://cg-taylor-acce.github.io/CG-Taylor/)]\n\n**POSE: Phased One-Step Adversarial Equilibrium for Video Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2508.21019)]\n[[Project](https://pose-paper.github.io/)]\n\n**Hyper-Bagel: A Unified Acceleration Framework for Multimodal Understanding and Generation** \\\n[[Website](https://arxiv.org/abs/2509.18824)]\n[[Project](https://hyper-bagel.github.io/)]\n\n**Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistency** \\\n[[Website](https://arxiv.org/abs/2510.08431)]\n[[Project](https://research.nvidia.com/labs/dir/rcm/)]\n\n**Self-Evaluation Unlocks Any-Step Text-to-Image Generation** \\\n[[Website](https://arxiv.org/abs/2512.22374)]\n[[Project](https://xinyu-andy.github.io/SelfE-project/)]\n\n**Transition Matching Distillation for Fast Video Generation** \\\n[[Website](https://arxiv.org/abs/2601.09881)]\n[[Project](https://research.nvidia.com/labs/genair/tmd/)]\n\n**Fast Autoregressive Video Diffusion and World Models with Temporal Cache Compression and Sparse Attention** \\\n[[Website](https://arxiv.org/abs/2602.01801)]\n[[Project](https://dvirsamuel.github.io/fast-auto-regressive-video/)]\n\n**FlashBlock: Attention Caching for Efficient Long-Context Block Diffusion** \\\n[[Website](https://arxiv.org/abs/2602.05305)]\n[[Project](https://caesarhhh.github.io/FlashBlock/)]\n\n**Adversarial Distribution Matching for Diffusion Distillation Towards Efficient Image and Video Synthesis** \\\n[[ICCV 2025 (Highlight)](https://arxiv.org/abs/2507.18569)]\n\n**OmniCache: A Trajectory-Oriented Global Perspective on Training-Free Cache Reuse for Diffusion Transformer Models** \\\n[[ICCV 2025](https://arxiv.org/abs/2508.16212)]\n\n**FasterDiT: Towards Faster Diffusion Transformers Training without Architecture Modification** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2410.10356)]\n\n**One-Step Diffusion Distillation through Score Implicit Matching** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2410.16794)]\n\n**Self-Corrected Flow Distillation for Consistent One-Step and Few-Step Text-to-Image Generation** \\\n[[AAAI 2025](https://arxiv.org/abs/2412.16906)]\n\n**TAP: A Token-Adaptive Predictor Framework for Training-Free Diffusion Acceleration** \\\n[[CVPR 2026](https://arxiv.org/abs/2603.03792)]\n\n**BlockDance: Reuse Structurally Similar Spatio-Temporal Features to Accelerate Diffusion Transformers** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.15927)]\n\n**PCM : Picard Consistency Model for Fast Parallel Sampling of Diffusion Models** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.19731)]\n\n**MVPortrait: Text-Guided Motion and Emotion Control for Multi-view Vivid Portrait Animation** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.19383)]\n\n**Revisiting Diffusion Models: From Generative Pre-training to One-Step Generation** \\\n[[ICML 2025](https://arxiv.org/abs/2506.09376)]\n\n**Accelerate High-Quality Diffusion Models with Inner Loop Feedback** \\\n[[Website](https://arxiv.org/abs/2501.13107)]\n\n**Accelerating Diffusion Transformer via Error-Optimized Cache** \\\n[[Website](https://arxiv.org/abs/2501.19243)]\n\n**DICE: Distilling Classifier-Free Guidance into Text Embeddings** \\\n[[Website](https://arxiv.org/abs/2502.03726)]\n\n**ProReflow: Progressive Reflow with Decomposed Velocity** \\\n[[Website](https://arxiv.org/abs/2503.04824)]\n\n**Inference-Time Diffusion Model Distillation** \\\n[[Website](https://arxiv.org/abs/2412.08871)]\n\n**Taming Consistency Distillation for Accelerated Human Image Animation** \\\n[[Website](https://arxiv.org/abs/2504.11143)]\n\n**Token Pruning for Caching Better: 9 Times Acceleration on Stable Diffusion for Free** \\\n[[Website](https://arxiv.org/abs/2501.00375)]\n\n**HarmoniCa: Harmonizing Training and Inference for Better Feature Cache in Diffusion Transformer Acceleration** \\\n[[Website](https://arxiv.org/abs/2410.01723)]\n\n**Diff-Instruct\\*: Towards Human-Preferred One-step Text-to-image Generative Models** \\\n[[Website](https://arxiv.org/abs/2410.20898)]\n\n**MLCM: Multistep Consistency Distillation of Latent Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2406.05768)]\n\n**EM Distillation for One-step Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2405.16852)]\n\n**AsymRnR: Video Diffusion Transformers Acceleration with Asymmetric Reduction and Restoration** \\\n[[Website](https://arxiv.org/abs/2412.11706)]\n\n**Score-of-Mixture Training: Training One-Step Generative Models Made Simple** \\\n[[Website](https://arxiv.org/abs/2502.09609)]\n\n**Partially Conditioned Patch Parallelism for Accelerated Diffusion Model Inference** \\\n[[Website](https://arxiv.org/abs/2412.02962)]\n\n**Importance-based Token Merging for Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2411.16720)]\n\n**Imagine Flash: Accelerating Emu Diffusion Models with Backward Distillation** \\\n[[Website](https://arxiv.org/abs/2405.05224)]\n\n**Accelerating Diffusion Models with One-to-Many Knowledge Distillation** \\\n[[Website](https://arxiv.org/abs/2410.04191)]\n\n**Accelerating Video Diffusion Models via Distribution Matching** \\\n[[Website](https://arxiv.org/abs/2412.05899)]\n\n**TDDSR: Single-Step Diffusion with Two Discriminators for Super Resolution** \\\n[[Website](https://arxiv.org/abs/2410.07663)]\n\n**DDIL: Improved Diffusion Distillation With Imitation Learning** \\\n[[Website](https://arxiv.org/abs/2410.11971)]\n\n**OSV: One Step is Enough for High-Quality Image to Video Generation** \\\n[[Website](https://arxiv.org/abs/2409.11367)]\n\n**Target-Driven Distillation: Consistency Distillation with Target Timestep Selection and Decoupled Guidance** \\\n[[Website](https://arxiv.org/abs/2409.01347)]\n\n**Token Caching for Diffusion Transformer Acceleration** \\\n[[Website](https://arxiv.org/abs/2409.18523)]\n\n**DiP-GO: A Diffusion Pruner via Few-step Gradient Optimization** \\\n[[Website](https://arxiv.org/abs/2410.16942)]\n\n**LazyDiT: Lazy Learning for the Acceleration of Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2412.12444)]\n\n**Flow Generator Matching** \\\n[[Website](https://arxiv.org/abs/2410.19310)]\n\n**Multistep Distillation of Diffusion Models via Moment Matching** \\\n[[Website](https://arxiv.org/abs/2406.04103)]\n\n**SFDDM: Single-fold Distillation for Diffusion models** \\\n[[Website](https://arxiv.org/abs/2405.14961)]\n\n**LAPTOP-Diff: Layer Pruning and Normalized Distillation for Compressing Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2404.11098)]\n\n**CogView3: Finer and Faster Text-to-Image Generation via Relay Diffusion** \\\n[[Website](https://arxiv.org/abs/2403.05121)]\n\n**SCott: Accelerating Diffusion Models with Stochastic Consistency Distillation** \\\n[[Website](https://arxiv.org/abs/2403.01505)]\n\n**Ditto: Accelerating Diffusion Model via Temporal Value Similarity** \\\n[[Website](https://arxiv.org/abs/2501.11211)]\n\n**Adaptive Non-Uniform Timestep Sampling for Diffusion Model Training** \\\n[[Website](https://arxiv.org/abs/2411.09998)]\n\n**TSD-SR: One-Step Diffusion with Target Score Distillation for Real-World Image Super-Resolution** \\\n[[Website](https://arxiv.org/abs/2411.18263)]\n\n**Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity** \\\n[[Website](https://arxiv.org/abs/2502.01776)]\n\n**Efficient Distillation of Classifier-Free Guidance using Adapters** \\\n[[Website](https://arxiv.org/abs/2503.07274)]\n\n**Denoising Score Distillation: From Noisy Diffusion Pretraining to One-Step High-Quality Generation** \\\n[[Website](https://arxiv.org/abs/2503.07578)]\n\n**Inductive Moment Matching** \\\n[[Website](https://arxiv.org/abs/2503.07565)]\n\n**High Quality Diffusion Distillation on a Single GPU with Relative and Absolute Position Matching** \\\n[[Website](https://arxiv.org/abs/2503.20744)]\n\n**DiTFastAttnV2: Head-wise Attention Compression for Multi-Modality Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2503.22796)]\n\n**Mean Flows for One-step Generative Modeling** \\\n[[Website](https://arxiv.org/abs/2505.13447)]\n\n**Faster Video Diffusion with Trainable Sparse Attention** \\\n[[Website](https://arxiv.org/abs/2505.13389)]\n\n**Accelerating Diffusion-based Super-Resolution with Dynamic Time-Spatial Sampling** \\\n[[Website](https://arxiv.org/abs/2505.12048)]\n\n**SRDiffusion: Accelerate Video Diffusion Inference via Sketching-Rendering Cooperation** \\\n[[Website](https://arxiv.org/abs/2505.19151)]\n\n**Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation** \\\n[[Website](https://arxiv.org/abs/2505.18875)]\n\n**RainFusion: Adaptive Video Generation Acceleration via Multi-Dimensional Visual Redundancy** \\\n[[Website](https://arxiv.org/abs/2505.21036)]\n\n**Foresight: Adaptive Layer Reuse for Accelerated and High-Quality Text-to-Video Generation** \\\n[[Website](https://arxiv.org/abs/2506.00329)]\n\n**Accelerating Diffusion Large Language Models with SlowFast: The Three Golden Principles** \\\n[[Website](https://arxiv.org/abs/2506.10848)]\n\n**Diffusion Transformer-to-Mamba Distillation for High-Resolution Image Generation** \\\n[[Website](https://arxiv.org/abs/2506.18999)]\n\n**Upsample What Matters: Region-Adaptive Latent Sampling for Accelerated Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2507.08422)]\n\n**Accelerating Parallel Diffusion Model Serving with Residual Compression** \\\n[[Website](https://arxiv.org/abs/2507.17511)]\n\n**SwiftVideo: A Unified Framework for Few-Step Video Generation through Trajectory-Distribution Alignment** \\\n[[Website](https://arxiv.org/abs/2508.06082)]\n\n**MixCache: Mixture-of-Cache for Video Diffusion Transformer Acceleration** \\\n[[Website](https://arxiv.org/abs/2508.12691)]\n\n**Forecast then Calibrate: Feature Caching as ODE for Efficient Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2508.16211)]\n\n**DiCache: Let Diffusion Model Determine Its Own Cache** \\\n[[Website](https://arxiv.org/abs/2508.17356)]\n\n**HiCache: Training-free Acceleration of Diffusion Models via Hermite Polynomial-based Feature Caching** \\\n[[Website](https://arxiv.org/abs/2508.16984)]\n\n**SpecDiff: Accelerating Diffusion Model Inference with Self-Speculation** \\\n[[Website](https://arxiv.org/abs/2509.13848)]\n\n**BWCache: Accelerating Video Diffusion Transformers through Block-Wise Caching** \\\n[[Website](https://arxiv.org/abs/2509.13789)]\n\n**RAPID^3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformer** \\\n[[Website](https://arxiv.org/abs/2509.22323)]\n\n**SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention** \\\n[[Website](https://arxiv.org/abs/2509.24006)]\n\n**CLQ: Cross-Layer Guided Orthogonal-based Quantization for Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2509.24416)]\n\n**Score Distillation of Flow Matching Models** \\\n[[Website](https://arxiv.org/abs/2509.25127)]\n\n**Let Features Decide Their Own Solvers: Hybrid Feature Caching for Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2510.04188)]\n\n**LinVideo: A Post-Training Framework towards O(n) Attention in Efficient Video Generation** \\\n[[Website](https://arxiv.org/abs/2510.08318)]\n\n**FreqCa: Accelerating Diffusion Models via Frequency-Aware Caching** \\\n[[Website](https://arxiv.org/abs/2510.08669)]\n\n**Hierarchical Koopman Diffusion: Fast Generation with Interpretable Diffusion Trajectory** \\\n[[Website](https://arxiv.org/abs/2510.12220)]\n\n**Test-Time Iterative Error Correction for Efficient Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2511.06250)]\n\n**From Structure to Detail: Hierarchical Distillation for Efficient Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2511.08930)]\n\n**PipeDiT: Accelerating Diffusion Transformers in Video Generation with Task Pipelining and Model Decoupling** \\\n[[Website](https://arxiv.org/abs/2511.12056)]\n\n**Flash-DMD: Towards High-Fidelity Few-Step Image Generation with Efficient Distillation and Joint Reinforcement Learning** \\\n[[Website](https://arxiv.org/abs/2511.20549)]\n\n**GalaxyDiT: Efficient Video Generation with Guidance Alignment and Adaptive Proxy in Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2512.03451)]\n\n**InvarDiff: Cross-Scale Invariance Caching for Accelerated Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2512.05134)]\n\n**USV: Unified Sparsification for Accelerating Video Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2512.05754)]\n\n**TwinFlow: Realizing One-step Generation on Large Models with Self-adversarial Flows** \\\n[[Website](https://arxiv.org/abs/2512.05150)]\n\n**Few-Step Distillation for Text-to-Image Generation: A Practical Guide** \\\n[[Website](https://arxiv.org/abs/2512.13006)]\n\n**On the Design of One-step Diffusion via Shortcutting Flow Paths** \\\n[[Website](https://arxiv.org/abs/2512.11831)]\n\n**OUSAC: Optimized Guidance Scheduling with Adaptive Caching for DiT Acceleration** \\\n[[Website](https://arxiv.org/abs/2512.14096)]\n\n**Plug-and-Play Fidelity Optimization for Diffusion Transformer Acceleration via Cumulative Error Minimization** \\\n[[Website](https://arxiv.org/abs/2512.23258)]\n\n**Forecast the Principal, Stabilize the Residual: Subspace-Aware Feature Caching for Efficient Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2601.07396)]\n\n**DisCa: Accelerating Video Diffusion Transformers with Distillation-Compatible Learnable Feature Caching** \\\n[[Website](https://arxiv.org/abs/2602.05449)]\n\n**NanoFLUX: Distillation-Driven Compression of Large Text-to-Image Generation Models for Mobile Devices** \\\n[[Website](https://arxiv.org/abs/2602.06879)]\n\n**DDiT: Dynamic Patch Scheduling for Efficient Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2602.16968)]\n\n**LESA: Learnable Stage-Aware Predictors for Diffusion Model Acceleration** \\\n[[Website](https://arxiv.org/abs/2602.20497)]\n\n**Analyzing and Improving Fast Sampling of Text-to-Image Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2603.00763)]\n\n**Adaptive Spectral Feature Forecasting for Diffusion Sampling Acceleration** \\\n[[Website](https://arxiv.org/abs/2603.01623)]\n\n**TC-Padé: Trajectory-Consistent Padé Approximation for Diffusion Acceleration** \\\n[[Website](https://arxiv.org/abs/2603.02943)]\n\n\n## Train-Free\n\n**AsyncDiff: Parallelizing Diffusion Models by Asynchronous Denoising** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2406.06911)]\n[[Project](https://czg1225.github.io/asyncdiff_page/)]\n[[Code](https://github.com/czg1225/AsyncDiff)]\n\n**Training-Free Adaptive Diffusion with Bounded Difference Approximation Strategy** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2410.09873)]\n[[Project](https://jiakangyuan.github.io/AdaptiveDiffusion-project-page/)]\n[[Code](https://github.com/UniModal4Reasoning/AdaptiveDiffusion)]\n\n\n**DeepCache: Accelerating Diffusion Models for Free** \\\n[[CVPR 2024](https://arxiv.org/abs/2312.00858)]\n[[Project](https://horseee.github.io/Diffusion_DeepCache/)]\n[[Code](https://github.com/horseee/DeepCache)]\n\n**Grouping First, Attending Smartly: Training-Free Acceleration for Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2505.14687)]\n[[Project](https://oliverrensu.github.io/project/GRAT/)]\n[[Code](https://github.com/OliverRensu/GRAT)]\n\n**Faster Diffusion: Rethinking the Role of the Encoder for Diffusion Model Inference** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2312.09608)]\n[[Code](https://github.com/hutaihang/faster-diffusion)]\n\n**DiTFastAttn: Attention Compression for Diffusion Transformer Models** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2406.08552)]\n[[Code](https://github.com/thu-nics/DiTFastAttn)]\n\n**Structural Pruning for Diffusion Models** \\\n[[NeurIPS 2023](https://arxiv.org/abs/2305.10924)]\n[[Code](https://github.com/VainF/Diff-Pruning)]\n\n**AutoDiffusion: Training-Free Optimization of Time Steps and Architectures for Automated Diffusion Model Acceleration** \\\n[[ICCV 2023](https://arxiv.org/abs/2309.10438)]\n[[Code](https://github.com/lilijiangg/AutoDiffusion)]\n\n**Agent Attention: On the Integration of Softmax and Linear Attention** \\\n[[ECCV 2024](https://arxiv.org/abs/2312.08874)]\n[[Code](https://github.com/LeapLabTHU/Agent-Attention)]\n\n**Attend to Not Attended: Structure-then-Detail Token Merging for Post-training DiT Acceleration** \\\n[[CVPR 2025](https://arxiv.org/abs/2505.11707)]\n[[Code](https://github.com/ICTMCG/SDTM)]\n\n**Token Merging for Fast Stable Diffusion** \\\n[[CVPRW 2024](https://arxiv.org/abs/2303.17604)]\n[[Code](https://github.com/dbolya/tomesd)]\n\n**LightCache: Memory-Efficient, Training-Free Acceleration for Video Generation** \\\n[[Website](https://arxiv.org/abs/2510.05367)]\n[[Code](https://github.com/NKUShaw/LightCache)]\n\n**FORA: Fast-Forward Caching in Diffusion Transformer Acceleration** \\\n[[Website](https://arxiv.org/abs/2407.01425)]\n[[Code](https://github.com/prathebaselva/FORA)]\n\n**Real-Time Video Generation with Pyramid Attention Broadcast** \\\n[[Website](https://arxiv.org/abs/2408.12588)]\n[[Code](https://github.com/NUS-HPC-AI-Lab/VideoSys)]\n\n**Accelerating Diffusion Transformers with Token-wise Feature Caching** \\\n[[Website](https://arxiv.org/abs/2410.05317)]\n[[Code](https://github.com/Shenyi-Z/ToCa)]\n\n**TGATE-V1: Cross-Attention Makes Inference Cumbersome in Text-to-Image Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2404.02747v1)]\n[[Code](https://github.com/HaozheLiu-ST/T-GATE)]\n\n**TGATE-V2: Faster Diffusion via Temporal Attention Decomposition** \\\n[[Website](https://arxiv.org/abs/2404.02747v2)]\n[[Code](https://github.com/HaozheLiu-ST/T-GATE)]\n\n**SmoothCache: A Universal Inference Acceleration Technique for Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2411.10510)]\n[[Code](https://github.com/Roblox/SmoothCache)]\n\n**Attention-Driven Training-Free Efficiency Enhancement of Diffusion Models** \\\n[[CVPR 2024](https://arxiv.org/abs/2405.05252)]\n[[Project](https://atedm.github.io/)]\n\n**Training-free Diffusion Acceleration with Bottleneck Sampling** \\\n[[Website](https://arxiv.org/abs/2503.18940)]\n[[Project](https://tyfeld.github.io/BottleneckSampling.github.io/)]\n\n**Cache Me if You Can: Accelerating Diffusion Models through Block Caching** \\\n[[Website](https://arxiv.org/abs/2312.03209)]\n[[Project](https://github.com/Shenyi-Z/ToCa)]\n\n**Fewer Denoising Steps or Cheaper Per-Step Inference: Towards Compute-Optimal Diffusion Model Deployment** \\\n[[ICCV 2025](https://arxiv.org/abs/2508.06160)]\n\n**Token Fusion: Bridging the Gap between Token Pruning and Token Merging** \\\n[[WACV 2024](https://arxiv.org/abs/2312.01026)]\n\n**Flexiffusion: Training-Free Segment-Wise Neural Architecture Search for Efficient Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2506.02488)]\n\n**PFDiff: Training-free Acceleration of Diffusion Models through the Gradient Guidance of Past and Future** \\\n[[Website](https://arxiv.org/abs/2408.08822)]\n\n**Δ-DiT: A Training-Free Acceleration Method Tailored for Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2406.01125)]\n\n**Adversarial Score identity Distillation: Rapidly Surpassing the Teacher in One Step** \\\n[[Website](https://arxiv.org/abs/2410.14919)]\n\n**Diff-Instruct++: Training One-step Text-to-image Generator Model to Align with Human Preferences** \\\n[[Website](https://arxiv.org/abs/2410.18881)]\n\n**Fast constrained sampling in pre-trained diffusion models** \\\n[[Website](https://arxiv.org/abs/2410.18804)]\n\n**Chipmunk: Training-Free Acceleration of Diffusion Transformers with Dynamic Column-Sparse Deltas** \\\n[[Website](https://arxiv.org/abs/2506.03275)]\n\n**ETC: training-free diffusion models acceleration with Error-aware Trend Consistency** \\\n[[Website](https://arxiv.org/abs/2510.24129)]\n\n\n## AR model\n\n**Distilled Decoding 1: One-step Sampling of Image Auto-regressive Models with Flow Matching** \\\n[[ICLR 2025](https://arxiv.org/abs/2412.17153)]\n[[Project](https://imagination-research.github.io/distilled-decoding/)]\n[[Code](https://github.com/imagination-research/distilled-decoding)]\n\n**Accelerating Auto-regressive Text-to-Image Generation with Training-free Speculative Jacobi Decoding** \\\n[[ICLR 2025](https://arxiv.org/abs/2410.01699)]\n[[Code](https://github.com/tyshiwo1/Accelerating-T2I-AR-with-SJD)]\n\n**LANTERN: Accelerating Visual Autoregressive Models with Relaxed Speculative Decoding** \\\n[[ICLR 2025](https://arxiv.org/abs/2410.03355)]\n[[Code](https://github.com/tyshiwo1/Accelerating-T2I-AR-with-SJD)]\n\n**Show-o Turbo: Towards Accelerated Unified Multimodal Understanding and Generation** \\\n[[Website](https://arxiv.org/abs/2502.05415)]\n[[Code](https://github.com/zhijie-group/Show-o-Turbo)]\n\n**SimpleAR: Pushing the Frontier of Autoregressive Visual Generation through Pretraining, SFT, and RL** \\\n[[Website](https://arxiv.org/abs/2504.11455)]\n[[Code](https://github.com/wdrink/SimpleAR)]\n\n**Speculative Jacobi-Denoising Decoding for Accelerating Autoregressive Text-to-image Generation** \\\n[[Website](https://arxiv.org/abs/2510.08994)]\n\n**SJD++: Improved Speculative Jacobi Decoding for Training-free Acceleration of Discrete Auto-regressive Text-to-Image Generation** \\\n[[Website](https://arxiv.org/abs/2512.07503)]\n\n**Hawk: Leveraging Spatial Context for Faster Autoregressive Text-to-Image Generation** \\\n[[Website](https://arxiv.org/abs/2510.25739)]\n\n**Fast-ARDiff: An Entropy-informed Acceleration Framework for Continuous Space Autoregressive Generation** \\\n[[Website](https://arxiv.org/abs/2512.08537)]\n\n\n\n## VAR model\n\n**Collaborative Decoding Makes Visual Auto-Regressive Modeling Efficient** \\\n[[CVPR 2025](https://arxiv.org/abs/2411.17787)]\n[[Project](https://czg1225.github.io/CoDe_page/)]\n[[Code](https://github.com/czg1225/CoDe)]\n\n**FastVAR: Linear Visual Autoregressive Modeling via Cached Token Pruning** \\\n[[ICCV 2025](https://arxiv.org/abs/2503.23367)]\n[[Project](https://fastvar.github.io/)]\n[[Code](https://github.com/csguoh/FastVAR)]\n\n**Memory-Efficient Visual Autoregressive Modeling with Scale-Aware KV Cache Compression** \\\n[[Website](https://arxiv.org/abs/2505.19602)]\n[[Code](https://github.com/StargazerX0/ScaleKV)]\n\n**SkipVAR: Accelerating Visual Autoregressive Modeling via Adaptive Frequency-Aware Skipping** \\\n[[Website](https://arxiv.org/abs/2506.08908)]\n[[Code](https://github.com/fakerone-li/SkipVAR)]\n\n**Frequency-Aware Autoregressive Modeling for Efficient High-Resolution Image Synthesis** \\\n[[Website](https://arxiv.org/abs/2507.20454v1)]\n[[Code](https://github.com/Caesarhhh/SparseVAR)]\n\n**LiteVAR: Compressing Visual Autoregressive Modelling with Efficient Attention and Quantization** \\\n[[Website](https://arxiv.org/abs/2411.17178)]\n\n\n# Image Restoration\n\n\n**Diffusion Prior-Based Amortized Variational Inference for Noisy Inverse Problems** \\\n[[ECCV 2024 Oral](https://arxiv.org/abs/2407.16125)]\n[[Project](https://mlvlab.github.io/DAVI-project/)]\n[[Code](https://github.com/mlvlab/DAVI)]\n\n\n**Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model** \\\n[[ICLR 2023 oral](https://arxiv.org/abs/2212.00490)]\n[[Project](https://wyhuai.github.io/ddnm.io/)]\n[[Code](https://github.com/wyhuai/DDNM)]\n\n**Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild** \\\n[[CVPR 2024](https://arxiv.org/abs/2401.13627)]\n[[Project](https://supir.xpixel.group/)]\n[[Code](https://github.com/Fanghua-Yu/SUPIR)]\n\n**Selective Hourglass Mapping for Universal Image Restoration Based on Diffusion Model** \\\n[[CVPR 2024](https://arxiv.org/abs/2403.11157)]\n[[Project](https://isee-laboratory.github.io/DiffUIR/)]\n[[Code](https://github.com/iSEE-Laboratory/DiffUIR)]\n\n**Zero-Reference Low-Light Enhancement via Physical Quadruple Priors** \\\n[[CVPR 2024](https://arxiv.org/abs/2403.12933)]\n[[Project](https://daooshee.github.io/QuadPrior-Website/)]\n[[Code](https://github.com/daooshee/QuadPrior/)]\n\n**From Posterior Sampling to Meaningful Diversity in Image Restoration** \\\n[[ICLR 2024](https://arxiv.org/abs/2310.16047)]\n[[Project](https://noa-cohen.github.io/MeaningfulDiversityInIR/)]\n[[Code](https://github.com/noa-cohen/MeaningfulDiversityInIR)]\n\n**Generative Diffusion Prior for Unified Image Restoration and Enhancement** \\\n[[CVPR 2023](https://arxiv.org/abs/2304.01247)]\n[[Project](https://generativediffusionprior.github.io/)]\n[[Code](https://github.com/Fayeben/GenerativeDiffusionPrior)]\n\n**MoE-DiffIR: Task-customized Diffusion Priors for Universal Compressed Image Restoration** \\\n[[ECCV 2024](https://arxiv.org/abs/2407.10833)]\n[[Project](https://renyulin-f.github.io/MoE-DiffIR.github.io/)]\n[[Code](https://github.com/renyulin-f/MoE-DiffIR)]\n\n**Image Restoration with Mean-Reverting Stochastic Differential Equations** \\\n[[ICML 2023](https://arxiv.org/abs/2301.11699)]\n[[Project](https://algolzw.github.io/ir-sde/index.html)]\n[[Code](https://github.com/Algolzw/image-restoration-sde)]\n\n**PhoCoLens: Photorealistic and Consistent Reconstruction in Lensless Imaging** \\\n[[NeurIPS 2024 Spotlight](https://arxiv.org/abs/2409.17996)]\n[[Project](https://phocolens.github.io/)]\n[[Code](https://github.com/PhoCoLens)]\n\n**Denoising Diffusion Models for Plug-and-Play Image Restoration** \\\n[[CVPR 2023 Workshop NTIRE](https://arxiv.org/abs/2305.08995)]\n[[Project](https://yuanzhi-zhu.github.io/DiffPIR/)]\n[[Code](https://github.com/yuanzhi-zhu/DiffPIR)]\n\n**FoundIR: Unleashing Million-scale Training Data to Advance Foundation Models for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2412.01427)]\n[[Project](https://foundir.net/)]\n[[Code](https://github.com/House-Leo/FoundIR)]\n\n**Improving Diffusion Inverse Problem Solving with Decoupled Noise Annealing** \\\n[[Website](https://arxiv.org/abs/2407.01521)]\n[[Project](https://daps-inverse-problem.github.io/)]\n[[Code](https://github.com/zhangbingliang2019/DAPS)]\n\n**SVFR: A Unified Framework for Generalized Video Face Restoration** \\\n[[Website](https://arxiv.org/abs/2501.01235)]\n[[Project](https://wangzhiyaoo.github.io/SVFR/)]\n[[Code](https://github.com/wangzhiyaoo/SVFR)]\n\n**DiffIR2VR-Zero: Zero-Shot Video Restoration with Diffusion-based Image Restoration Models** \\\n[[Website](https://arxiv.org/abs/2407.01519)]\n[[Project](https://jimmycv07.github.io/DiffIR2VR_web/)]\n[[Code](https://github.com/jimmycv07/DiffIR2VR-Zero)]\n\n**Solving Video Inverse Problems Using Image Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2409.02574)]\n[[Project](https://solving-video-inverse.github.io/main/)]\n[[Code](https://github.com/solving-video-inverse/codes)]\n\n**RestoreVAR: Visual Autoregressive Generation for All-in-One Image Restoration** \\\n[[Website](https://arxiv.org/abs/2505.18047)]\n[[Project](https://sudraj2002.github.io/restorevarpage/)]\n[[Code](https://github.com/sudraj2002/RestoreVAR)]\n\n**Learning Efficient and Effective Trajectories for Differential Equation-based Image Restoration** \\\n[[Website](https://arxiv.org/abs/2410.04811)]\n[[Project](https://zhu-zhiyu.github.io/FLUX-IR/)]\n[[Code](https://github.com/ZHU-Zhiyu/FLUX-IR)]\n\n**GenDR: Lightning Generative Detail Restorator** \\\n[[Website](https://arxiv.org/abs/2503.06790)]\n[[Project](https://icandle.github.io/gendr_page/)]\n[[Code](https://github.com/icandle/GenDR)]\n\n**AutoDIR: Automatic All-in-One Image Restoration with Latent Diffusion** \\\n[[Website](https://arxiv.org/abs/2310.10123)]\n[[Project](https://jiangyitong.github.io/AutoDIR_webpage/)]\n[[Code](https://github.com/jiangyitong/AutoDIR)]\n\n**SeedVR2: One-Step Video Restoration via Diffusion Adversarial Post-Training** \\\n[[Website](https://arxiv.org/abs/2506.05301)]\n[[Project](https://iceclear.github.io/projects/seedvr2/)]\n[[Code](https://github.com/IceClear/SeedVR2)]\n\n**Text-Aware Image Restoration with Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2506.09993)]\n[[Project](https://cvlab-kaist.github.io/TAIR/)]\n[[Code](https://github.com/cvlab-kaist/TAIR)]\n\n**LucidFlux: Caption-Free Universal Image Restoration via a Large-Scale Diffusion Transformer** \\\n[[Website](https://arxiv.org/abs/2509.22414)]\n[[Project](https://w2genai-lab.github.io/LucidFlux/)]\n[[Code](https://github.com/W2GenAI-Lab/LucidFlux)]\n\n**Zero-Shot Video Deraining with Video Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2511.18537)]\n[[Project](https://tvaranka.github.io/ZSVD/)]\n[[Code](https://github.com/tvaranka/ZSVD)]\n\n**TPGDiff: Hierarchical Triple-Prior Guided Diffusion for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2601.20306)]\n[[Project](https://leoyjtu.github.io/tpgdiff-project/)]\n[[Code](https://github.com/leoyjTu/TPGDiff)]\n\n**FlowIE: Efficient Image Enhancement via Rectified Flow** \\\n[[CVPR 2024 oral](https://arxiv.org/abs/2406.00508)]\n[[Code](https://github.com/EternalEvan/FlowIE)]\n\n**ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting** \\\n[[NeurIPS 2023 (Spotlight)](https://arxiv.org/abs/2307.12348)]\n[[Code](https://github.com/zsyOAOA/ResShift)]\n\n**GibbsDDRM: A Partially Collapsed Gibbs Sampler for Solving Blind Inverse Problems with Denoising Diffusion Restoration** \\\n[[ICML 2023 oral](https://arxiv.org/abs/2301.12686)]\n[[Code](https://github.com/sony/gibbsddrm)]\n\n**Diffusion Priors for Variational Likelihood Estimation and Image Denoising** \\\n[[NeurIPS 2024 Spotlight](https://arxiv.org/abs/2410.17521)]\n[[Code](https://github.com/HUST-Tan/DiffusionVI)]\n\n**DiffIR: Efficient Diffusion Model for Image Restoration**\\\n\u003c!-- [[ICCV 2023](https://openaccess.thecvf.com/content/ICCV2023/papers/Xia_DiffIR_Efficient_Diffusion_Model_for_Image_Restoration_ICCV_2023_paper.pdf)] --\u003e\n[[ICCV 2023](https://arxiv.org/abs/2303.09472)] \n[[Code](https://github.com/Zj-BinXia/DiffIR)]\n\n**Compression-Aware One-Step Diffusion Model for JPEG Artifact Removal** \\\n[[ICCV 2025](https://arxiv.org/pdf/2502.09873)]\n[[Code](https://github.com/jp-guo/CODiff)]\n\n**Image Restoration by Denoising Diffusion Models with Iteratively Preconditioned Guidance** \\\n[[CVPR 2024](https://arxiv.org/abs/2312.16519)]\n[[Code](https://github.com/tirer-lab/DDPG)]\n\n**InstaRevive: One-Step Image Enhancement via Dynamic Score Matching** \\\n[[ICLR 2025](https://arxiv.org/abs/2504.15513)]\n[[Code](https://github.com/EternalEvan/InstaRevive)]\n\n**LightenDiffusion: Unsupervised Low-Light Image Enhancement with Latent-Retinex Diffusion Models** \\\n[[ECCV 2024](https://arxiv.org/abs/2407.08939)]\n[[Code](https://github.com/JianghaiSCU/LightenDiffusion)]\n\n**Rethinking Video Deblurring with Wavelet-Aware Dynamic Transformer and Diffusion Model** \\\n[[ECCV 2024](https://arxiv.org/abs/2408.13459)]\n[[Code](https://github.com/Chen-Rao/VD-Diff)]\n\n**DAVI: Diffusion Prior-Based Amortized Variational Inference for Noisy Inverse Problem** \\\n[[ECCV 2024](https://arxiv.org/abs/2407.16125)]\n[[Code](https://github.com/mlvlab/DAVI)]\n\n**Low-Light Image Enhancement with Wavelet-based Diffusion Models** \\\n[[SIGGRAPH Asia 2023](https://arxiv.org/abs/2306.00306)]\n[[Code](https://github.com/JianghaiSCU/Diffusion-Low-Light)]\n\n**Residual Denoising Diffusion Models** \\\n[[CVPR 2024](https://arxiv.org/abs/2308.13712)]\n[[Code](https://github.com/nachifur/RDDM)]\n\n**Diff-Plugin: Revitalizing Details for Diffusion-based Low-level Tasks** \\\n[[CVPR 2024](https://arxiv.org/abs/2403.00644)]\n[[Code](https://github.com/yuhaoliu7456/Diff-Plugin)]\n\n**Learning Hazing to Dehazing: Towards Realistic Haze Generation for Real-World Image Dehazing** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.19262)]\n[[Code](https://github.com/ruiyi-w/Learning-Hazing-to-Dehazing)]\n\n**Deep Equilibrium Diffusion Restoration with Parallel Sampling** \\\n[[CVPR 2024](https://arxiv.org/abs/2311.11600)]\n[[Code](https://github.com/caojiezhang/deqir)]\n\n**Unleashing the Potential of the Semantic Latent Space in Diffusion Models for Image Dehazing** \\\n[[ECCV 2024](https://arxiv.org/abs/2509.20091)]\n[[Code](https://github.com/aaaasan111/difflid)]\n\n**An Expectation-Maximization Algorithm for Training Clean Diffusion Models from Corrupted Observations** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2407.01014)]\n[[Code](https://github.com/weiminbai/EMDiffusion)]\n\n**ReFIR: Grounding Large Restoration Models with Retrieval Augmentation** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2410.05601)]\n[[Code](https://github.com/csguoh/ReFIR)]\n\n**DreamClear: High-Capacity Real-World Image Restoration with Privacy-Safe Dataset Curation** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2410.18666)]\n[[Code](https://github.com/shallowdream204/DreamClear)]\n\n**Reconciling Stochastic and Deterministic Strategies for Zero-shot Image Restoration using Diffusion Model in Dual** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.01288)]\n[[Code](https://github.com/ChongWang1024/RDMD)]\n\n**Learning to See in the Extremely Dark** \\\n[[ICCV 2025](https://arxiv.org/abs/2506.21132)]\n[[Code](https://github.com/JianghaiSCU/SIED)]\n\n**Exploiting Diffusion Prior for Real-World Image Dehazing with Unpaired Training** \\\n[[AAAI 2025](https://arxiv.org/abs/2503.15017)]\n[[Code](https://github.com/ywxjm/Diff-Dehazer)]\n\n**Seeing Through the Rain: Resolving High-Frequency Conflicts in Deraining and Super-Resolution via Diffusion Guidance** \\\n[[AAAI 2026](https://arxiv.org/abs/2511.12419)]\n[[Code](https://github.com/PRIS-CV/DHGM)]\n\n**Genuine Knowledge from Practice: Diffusion Test-Time Adaptation for Video Adverse Weather Removal** \\\n[[CVPR 2024](https://arxiv.org/abs/2403.07684)]\n[[Code](https://github.com/scott-yjyang/DiffTTA)]\n\n**Enhancing Diffusion Model Stability for Image Restoration via Gradient Management** \\\n[[ACM MM 2025](https://arxiv.org/abs/2507.06656)]\n[[Code](https://github.com/74587887/SPGD)]\n\n**PerTouch: VLM-Driven Agent for Personalized and Semantic Image Retouching** \\\n[[AAAI 2026](https://arxiv.org/abs/2511.12998)]\n[[Code](https://github.com/Auroral703/PerTouch)]\n\n**Refusion: Enabling Large-Size Realistic Image Restoration with Latent-Space Diffusion Models** \\\n[[CVPR 2023 Workshop NTIRE](https://arxiv.org/abs/2304.08291)]\n[[Code](https://github.com/Algolzw/image-restoration-sde)]\n\n**Equipping Diffusion Models with Differentiable Spatial Entropy for Low-Light Image Enhancement** \\\n[[CVPR 2024 Workshop NTIRE](https://arxiv.org/abs/2404.09735)]\n[[Code](https://github.com/shermanlian/spatial-entropy-loss)]\n\n**JPEG Artifact Correction using Denoising Diffusion Restoration Models** \\\n[[NeurIPS 2022 Workshop](https://arxiv.org/abs/2209.11888)]\n[[Code](https://github.com/bahjat-kawar/ddrm-jpeg)]\n\n**FlowDPS: Flow-Driven Posterior Sampling for Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2503.08136)]\n[[Code](https://github.com/FlowDPS-Inverse/FlowDPS)]\n\n**InstructRestore: Region-Customized Image Restoration with Human Instructions** \\\n[[Website](https://arxiv.org/abs/2503.24357)]\n[[Code](https://github.com/shuaizhengliu/InstructRestore)]\n\n**Decoupled Data Consistency with Diffusion Purification for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2403.06054)]\n[[Code](https://github.com/morefre/decoupled-data-consistency-with-diffusion-purification-for-image-restoration)]\n\n\n**One-Step Diffusion Model for Image Motion-Deblurring** \\\n[[Website](https://arxiv.org/abs/2503.06537)]\n[[Code](https://github.com/xyLiu339/OSDD)]\n\n**Reversing the Damage: A QP-Aware Transformer-Diffusion Approach for 8K Video Restoration under Codec Compression** \\\n[[Website](https://arxiv.org/abs/2412.08912)]\n[[Code](https://github.com/alimd94/DiQP)]\n\n**Zero-Shot Adaptation for Approximate Posterior Sampling of Diffusion Models in Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2407.11288)]\n[[Code](https://github.com/ualcalar17/ZAPS)]\n\n**Improving Diffusion-based Inverse Algorithms under Few-Step Constraint via Learnable Linear Extrapolation** \\\n[[Website](https://arxiv.org/abs/2503.10103)]\n[[Code](https://github.com/weigerzan/LLE_inverse_problem)]\n\n**DGSolver: Diffusion Generalist Solver with Universal Posterior Sampling for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2504.21487)]\n[[Code](https://github.com/MiliLab/DGSolver)]\n\n**DeblurDiff: Real-World Image Deblurring with Generative Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2502.03810)]\n[[Code](https://github.com/kkkls/DeblurDiff)]\n\n**UniProcessor: A Text-induced Unified Low-level Image Processor** \\\n[[Website](https://arxiv.org/abs/2407.20928)]\n[[Code](https://github.com/IntMeGroup/UniProcessor)]\n\n**Zero-Shot Image Restoration Using Few-Step Guidance of Consistency Models (and Beyond)** \\\n[[Website](https://arxiv.org/abs/2412.20596)]\n[[Code](https://github.com/tirer-lab/CM4IR)]\n\n**Varformer: Adapting VAR's Generative Prior for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2412.21063)]\n[[Code](https://github.com/siywang541/Varformer)]\n\n**Low-Light Image Enhancement via Generative Perceptual Priors** \\\n[[Website](https://arxiv.org/abs/2412.20916)]\n[[Code](https://github.com/LowLevelAI/GPP-LLIE)]\n\n**PnP-Flow: Plug-and-Play Image Restoration with Flow Matching** \\\n[[Website](https://arxiv.org/abs/2410.02423)]\n[[Code](https://github.com/annegnx/PnP-Flow)]\n\n**VIIS: Visible and Infrared Information Synthesis for Severe Low-light Image Enhancement** \\\n[[Website](https://arxiv.org/abs/2412.13655)]\n[[Code](https://github.com/Chenz418/VIIS)]\n\n**Deep Data Consistency: a Fast and Robust Diffusion Model-based Solver for Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2405.10748)]\n[[Code](https://github.com/Hanyu-Chen373/DeepDataConsistency)]\n\n**Learning A Coarse-to-Fine Diffusion Transformer for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2308.08730)]\n[[Code](https://github.com/wlydlut/C2F-DFT)]\n\n**ExpRDiff: Short-exposure Guided Diffusion Model for Realistic Local Motion Deblurring** \\\n[[Website](https://arxiv.org/abs/2412.09193)]\n[[Code](https://github.com/yzb1997/ExpRDiff)]\n\n**Stimulating the Diffusion Model for Image Denoising via Adaptive Embedding and Ensembling** \\\n[[Website](https://arxiv.org/abs/2307.03992)]\n[[Code](https://github.com/Li-Tong-621/DMID)]\n\n**Solving Linear Inverse Problems Provably via Posterior Sampling with Latent Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2307.00619)]\n[[Code](https://github.com/liturout/psld)]\n\n**Sagiri: Low Dynamic Range Image Enhancement with Generative Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2406.09389)]\n[[Code](https://github.com/ztMotaLee/Sagiri)]\n\n**Frequency Compensated Diffusion Model for Real-scene Dehazing** \\\n[[Website](https://arxiv.org/abs/2308.10510)]\n[[Code](https://github.com/W-Jilly/frequency-compensated-diffusion-model-pytorch)]\n\n**Efficient Image Deblurring Networks based on Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2401.05907)]\n[[Code](https://github.com/bnm6900030/swintormer)]\n\n**Blind Image Restoration via Fast Diffusion Inversion** \\\n[[Website](https://arxiv.org/abs/2405.19572)]\n[[Code](https://github.com/hamadichihaoui/BIRD)]\n\n**DMPlug: A Plug-in Method for Solving Inverse Problems with Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2405.16749)]\n[[Code](https://github.com/sun-umn/DMPlug)]\n\n**Accelerating Diffusion Models for Inverse Problems through Shortcut Sampling** \\\n[[Website](https://arxiv.org/abs/2305.16965)]\n[[Code](https://github.com/GongyeLiu/SSD)]\n\n**Denoising as Adaptation: Noise-Space Domain Adaptation for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2406.18516)]\n[[Code](https://github.com/KangLiao929/Noise-DA/)]\n\n**Unlimited-Size Diffusion Restoration** \\\n[[Website](https://arxiv.org/abs/2303.00354)]\n[[Code](https://github.com/wyhuai/DDNM/tree/main/hq_demo)]\n\n**UniDB++: Fast Sampling of Unified Diffusion Bridge** \\\n[[Website](https://arxiv.org/abs/2505.21528)]\n[[Code](https://github.com/2769433owo/UniDB-plusplus)]\n\n**VmambaIR: Visual State Space Model for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2403.11423)]\n[[Code](https://github.com/AlphacatPlus/VmambaIR)]\n\n**InvFussion: Bridging Supervised and Zero-shot Diffusion for Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2504.01689)]\n[[Code](https://github.com/noamelata/InvFussion)]\n\n**Using diffusion model as constraint: Empower Image Restoration Network Training with Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2406.19030)]\n[[Code](https://github.com/JosephTiTan/DiffLoss)]\n\n**Super-resolving Real-world Image Illumination Enhancement: A New Dataset and A Conditional Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2410.12961)]\n[[Code](https://github.com/Yaofang-Liu/Super-Resolving)]\n\n**BD-Diff: Generative Diffusion Model for Image Deblurring on Unknown Domains with Blur-Decoupled Learning** \\\n[[Website](https://arxiv.org/abs/2502.01522)]\n[[Code](https://github.com/donahowe/BD-Diff)]\n\n**IRBridge: Solving Image Restoration Bridge with Pre-trained Generative Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2505.24406)]\n[[Code](https://github.com/HashWang-null/IRBridge)]\n\n**Degradation-Consistent Learning via Bidirectional Diffusion for Low-Light Image Enhancement** \\\n[[Website](https://arxiv.org/abs/2507.18144)]\n[[Code](https://github.com/hejh8/BidDiff)]\n\n**Residual Diffusion Bridge Model for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2510.23116)]\n[[Code](https://github.com/MiliLab/RDBM)]\n\n**Learnable Fractional Reaction-Diffusion Dynamics for Under-Display ToF Imaging and Beyond** \\\n[[Website](https://arxiv.org/abs/2511.01704)]\n[[Code](https://github.com/wudiqx106/LFRD2)]\n\n**EndoIR: Degradation-Agnostic All-in-One Endoscopic Image Restoration via Noise-Aware Routing Diffusion** \\\n[[Website](https://arxiv.org/abs/2511.05873)]\n[[Code](https://github.com/DavisMeee/EndoIR)]\n\n**Equivariant Sampling for Improving Diffusion Model-based Image Restoration** \\\n[[Website](https://arxiv.org/abs/2511.09965)]\n[[Code](https://github.com/FouierL/EquS)]\n\n**Fose: Fusion of One-Step Diffusion and End-to-End Network for Pansharpening** \\\n[[Website](https://arxiv.org/abs/2512.17202)]\n[[Code](https://github.com/Kai-Liu001/Fose)]\n\n**MiM-DiT: MoE in MoE with Diffusion Transformers for All-in-One Image Restoration** \\\n[[Website](https://arxiv.org/abs/2603.02710)]\n[[Code](https://github.com/kkkls/MIM-DiT)]\n\n**Toward Generalized Image Quality Assessment: Relaxing the Perfect Reference Quality Assumption** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.11221)]\n[[Project](https://tianhewu.github.io/A-FINE-page.github.io/)]\n\n**TIP: Text-Driven Image Processing with Semantic and Restoration Instructions** \\\n[[ECCV 2024](https://arxiv.org/abs/2312.11595)]\n[[Project](https://chenyangqiqi.github.io/tip/)]\n\n**Warped Diffusion: Solving Video Inverse Problems with Image Diffusion Models** \\\n[[NeurIPS 2024](https://arxiv.org/abs/2410.16152)]\n[[Project](https://giannisdaras.github.io/warped_diffusion.github.io/)]\n\n**GenDeg: Diffusion-Based Degradation Synthesis for Generalizable All-in-One Image Restoration** \\\n[[Website](https://arxiv.org/abs/2411.17687)]\n[[Project](https://sudraj2002.github.io/gendegpage/)]\n\n**VISION-XL: High Definition Video Inverse Problem Solver using Latent Image Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2412.00156)]\n[[Project](https://vision-xl.github.io/)]\n\n**SeedVR: Seeding Infinity in Diffusion Transformer Towards Generic Video Restoration** \\\n[[Website](https://arxiv.org/abs/2501.01320)]\n[[Project](https://iceclear.github.io/projects/seedvr/)]\n\n**SILO: Solving Inverse Problems with Latent Operators** \\\n[[Website](https://arxiv.org/abs/2501.11746)]\n[[Project](https://ronraphaeli.github.io/SILO-website/)]\n\n**Proxies for Distortion and Consistency with Applications for Real-World Image Restoration** \\\n[[Website](https://arxiv.org/abs/2501.12102)]\n[[Project](https://man-sean.github.io/elad-website/)]\n\n**UniCoRN: Latent Diffusion-based Unified Controllable Image Restoration Network across Multiple Degradations** \\\n[[Website](https://arxiv.org/abs/2503.15868)]\n[[Project](https://codejaeger.github.io/unicorn-gh/)]\n\n**Lumina-OmniLV: A Unified Multimodal Framework for General Low-Level Vision** \\\n[[Website](https://arxiv.org/abs/2504.04903)]\n[[Project](https://andrew0613.github.io/OmniLV_page/)]\n\n**From Events to Clarity: The Event-Guided Diffusion Framework for Dehazing** \\\n[[Website](https://arxiv.org/abs/2511.11944)]\n[[Project](https://evdehaze.github.io/)]\n\n**FlowSteer: Conditioning Flow Field for Consistent Image Restoration** \\\n[[Website](https://arxiv.org/abs/2512.08125)]\n[[Project](https://tharindu-nirmal.github.io/FlowSteer/)]\n\n**CreativeVR: Diffusion-Prior-Guided Approach for Structure and Motion Restoration in Generative and Real Videos** \\\n[[Website](https://arxiv.org/abs/2512.12060)]\n[[Project](https://daveishan.github.io/creativevr-webpage/)]\n\n**BlurDM: A Blur Diffusion Model for Image Deblurring** \\\n[[NeurIPS 2025](https://arxiv.org/abs/2512.03979)]\n\n**DREAMCLEAN: RESTORING CLEAN IMAGE USING DEEP DIFFUSION PRIOR** \\\n[[ICLR 2025](https://openreview.net/forum?id=6ALuy19mPa)]\n\n**Diff-Retinex: Rethinking Low-light Image Enhancement with A Generative Diffusion Model** \\\n[[ICCV 2023](https://arxiv.org/abs/2308.13164)]\n\n**Multiscale Structure Guided Diffusion for Image Deblurring** \\\n[[ICCV 2023](https://arxiv.org/abs/2212.01789)]\n\n**Boosting Image Restoration via Priors from Pre-trained Models** \\\n[[CVPR 2024](https://arxiv.org/abs/2403.06793)]\n\n**Acquire and then Adapt: Squeezing out Text-to-Image Model for Image Restoration** \\\n[[CVPR 2025](https://arxiv.org/abs/2504.15159)]\n\n**Visual-Instructed Degradation Diffusion for All-in-One Image Restoration** \\\n[[CVPR 2025](https://arxiv.org/abs/2506.16960)]\n\n**Reversing Flow for Image Restoration** \\\n[[CVPR 2025](https://arxiv.org/abs/2506.16961)]\n\n**Dual Prompting Image Restoration with Diffusion Transformers** \\\n[[CVPR 2025](https://arxiv.org/abs/2504.17825)]\n\n**Integrating Intermediate Layer Optimization and Projected Gradient Descent for Solving Inverse Problems with Diffusion Models** \\\n[[ICML 2025](https://arxiv.org/abs/2505.20789)]\n\n**Exploiting Diffusion Prior for Task-driven Image Restoration** \\\n[[ICCV 2025](https://arxiv.org/abs/2507.22459)]\n\n**Seeing Beyond Haze: Generative Nighttime Image Dehazing** \\\n[[Website](https://arxiv.org/abs/2503.08073)]\n\n**Inverse Problem Sampling in Latent Space Using Sequential Monte Carlo** \\\n[[Website](https://arxiv.org/abs/2502.05908)]\n\n**Human Body Restoration with One-Step Diffusion Model and A New Benchmark** \\\n[[Website](https://arxiv.org/abs/2502.01411)]\n\n**A Modular Conditional Diffusion Framework for Image Reconstruction** \\\n[[Website](https://arxiv.org/abs/2411.05993)]\n\n**Solving Inverse Problems using Diffusion with Fast Iterative Renoising** \\\n[[Website](https://arxiv.org/abs/2501.17468)]\n\n**Unpaired Photo-realistic Image Deraining with Energy-informed Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2407.17193)]\n\n**Particle-Filtering-based Latent Diffusion for Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2408.13868)]\n\n**Bayesian Conditioned Diffusion Models for Inverse Problem** \\\n[[Website](https://arxiv.org/abs/2406.09768)]\n\n**ReCo-Diff: Explore Retinex-Based Condition Strategy in Diffusion Model for Low-Light Image Enhancement** \\\n[[Website](https://arxiv.org/abs/2312.12826)]\n\n**Multimodal Prompt Perceiver: Empower Adaptiveness, Generalizability and Fidelity for All-in-One Image Restoration** \\\n[[Website](https://arxiv.org/abs/2312.02918)]\n\n**Tell Me What You See: Text-Guided Real-World Image Denoising**\\\n[[Website](https://arxiv.org/abs/2312.10191)]\n\n**Zero-LED: Zero-Reference Lighting Estimation Diffusion Model for Low-Light Image Enhancement** \\\n[[Website](https://arxiv.org/abs/2403.02879)]\n\n**Prototype Clustered Diffusion Models for Versatile Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2407.09768)]\n\n**AGLLDiff: Guiding Diffusion Models Towards Unsupervised Training-free Real-world Low-light Image Enhancement** \\\n[[Website](https://arxiv.org/abs/2407.14900)]\n\n**Taming Generative Diffusion for Universal Blind Image Restoration** \\\n[[Website](https://arxiv.org/abs/2408.11287)]\n\n**Efficient Image Restoration through Low-Rank Adaptation and Stable Diffusion XL** \\\n[[Website](https://arxiv.org/abs/2408.17060)]\n\n**TDM: Temporally-Consistent Diffusion Model for All-in-One Real-World Video Restoration** \\\n[[Website](https://arxiv.org/abs/2501.02269)]\n\n**Empirical Bayesian image restoration by Langevin sampling with a denoising diffusion implicit prior** \\\n[[Website](https://arxiv.org/abs/2409.04384)]\n\n**Enhancing Diffusion Models for Inverse Problems with Covariance-Aware Posterior Sampling** \\\n[[Website](https://arxiv.org/abs/2412.20045)]\n\n**Data-free Distillation with Degradation-prompt Diffusion for Multi-weather Image Restoration** \\\n[[Website](https://arxiv.org/abs/2409.03455)]\n\n**FreeEnhance: Tuning-Free Image Enhancement via Content-Consistent Noising-and-Denoising Process** \\\n[[Website](https://arxiv.org/abs/2409.07451)]\n\n**Diffusion State-Guided Projected Gradient for Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2410.03463)]\n\n**InstantIR: Blind Image Restoration with Instant Generative Reference** \\\n[[Website](https://arxiv.org/abs/2410.06551)]\n\n**Score-Based Variational Inference for Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2410.05646)]\n\n**Towards Flexible and Efficient Diffusion Low Light Enhancer** \\\n[[Website](https://arxiv.org/abs/2410.12346)]\n\n**AdaQual-Diff: Diffusion-Based Image Restoration via Adaptive Quality Prompting** \\\n[[Website](https://arxiv.org/abs/2504.12605)]\n\n**G2D2: Gradient-guided Discrete Diffusion for image inverse problem solving** \\\n[[Website](https://arxiv.org/abs/2410.14710)]\n\n**AllRestorer: All-in-One Transformer for Image Restoration under Composite Degradations** \\\n[[Website](https://arxiv.org/abs/2411.10708)]\n\n**STeP: A General and Scalable Framework for Solving Video Inverse Problems with Spatiotemporal Diffusion Priors** \\\n[[Website](https://arxiv.org/abs/2504.07549)]\n\n**DiffMVR: Diffusion-based Automated Multi-Guidance Video Restoration** \\\n[[Website](https://arxiv.org/abs/2411.18745)]\n\n**Blind Inverse Problem Solving Made Easy by Text-to-Image Latent Diffusion** \\\n[[Website](https://arxiv.org/abs/2412.00557)]\n\n**DIVD: Deblurring with Improved Video Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2412.00773)]\n\n**Beyond Pixels: Text Enhances Generalization in Real-World Image Restoration** \\\n[[Website](https://arxiv.org/abs/2412.00878)]\n\n**Enhancing and Accelerating Diffusion-Based Inverse Problem Solving through Measurements Optimization** \\\n[[Website](https://arxiv.org/abs/2412.03941)]\n\n**Are Conditional Latent Diffusion Models Effective for Image Restoration?** \\\n[[Website](https://arxiv.org/abs/2412.09324)]\n\n**Consistent Diffusion: Denoising Diffusion Model with Data-Consistent Training for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2412.12550)]\n\n**DiffStereo: High-Frequency Aware Diffusion Model for Stereo Image Restoration** \\\n[[Website](https://arxiv.org/abs/2501.10325)]\n\n**Diffusion Restoration Adapter for Real-World Image Restoration** \\\n[[Website](https://arxiv.org/abs/2502.20679)]\n\n**Noise Synthesis for Low-Light Image Denoising with Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2503.11262)]\n\n**A Simple Combination of Diffusion Models for Better Quality Trade-Offs in Image Denoising** \\\n[[Website](https://arxiv.org/abs/2503.14654)]\n\n**Temporal-Consistent Video Restoration with Pre-trained Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2503.14863)]\n\n**Diffusion Image Prior** \\\n[[Website](https://arxiv.org/abs/2503.21410)]\n\n**Invert2Restore: Zero-Shot Degradation-Blind Image Restoration** \\\n[[Website](https://arxiv.org/abs/2503.21486)]\n\n**Blind Inversion using Latent Diffusion Priors** \\\n[[Website](https://arxiv.org/abs/2407.01027)]\n\n**CDI: Blind Image Restoration Fidelity Evaluation based on Consistency with Degraded Image** \\\n[[Website](https://arxiv.org/abs/2501.14264)]\n\n**IDDM: Bridging Synthetic-to-Real Domain Gap from Physics-Guided Diffusion for Real-world Image Dehazing** \\\n[[Website](https://arxiv.org/abs/2504.21385)]\n\n**DiffVQA: Video Quality Assessment Using Diffusion Feature Extractor** \\\n[[Website](https://arxiv.org/abs/2505.03261)]\n\n**LatentINDIGO: An INN-Guided Latent Diffusion Algorithm for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2505.12935)]\n\n**Dual Ascent Diffusion for Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2505.17353)]\n\n**Restoring Real-World Images with an Internal Detail Enhancement Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2505.18674)]\n\n**HAODiff: Human-Aware One-Step Diffusion via Dual-Prompt Guidance** \\\n[[Website](https://arxiv.org/abs/2505.19742)]\n\n**DarkDiff: Advancing Low-Light Raw Enhancement by Retasking Diffusion Models for Camera ISP** \\\n[[Website](https://arxiv.org/abs/2505.23743)]\n\n**Latent Guidance in Diffusion Models for Perceptual Evaluations** \\\n[[Website](https://arxiv.org/abs/2506.00327)]\n\n**Solving Inverse Problems with FLAIR** \\\n[[Website](https://arxiv.org/abs/2506.02680)]\n\n**Solving Inverse Problems via Diffusion-Based Priors: An Approximation-Free Ensemble Sampling Approach** \\\n[[Website](https://arxiv.org/abs/2506.03979)]\n\n**Restereo: Diffusion stereo video generation and restoration** \\\n[[Website](https://arxiv.org/abs/2506.06023)]\n\n**UniRes: Universal Image Restoration for Complex Degradations** \\\n[[Website](https://arxiv.org/abs/2506.05599)]\n\n**Zero-Shot Solving of Imaging Inverse Problems via Noise-Refined Likelihood Guided Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2506.13391)]\n\n**Elucidating and Endowing the Diffusion Training Paradigm for General Image Restoration** \\\n[[Website](https://arxiv.org/abs/2506.21722)]\n\n**LD-RPS: Zero-Shot Unified Image Restoration via Latent Diffusion Recurrent Posterior Sampling** \\\n[[Website](https://arxiv.org/abs/2507.00790)]\n\n**Harnessing Diffusion-Yielded Score Priors for Image Restoration** \\\n[[Website](https://arxiv.org/abs/2507.20590)]\n\n**UniLDiff: Unlocking the Power of Diffusion Priors for All-in-One Image Restoration** \\\n[[Website](https://arxiv.org/abs/2507.23685)]\n\n**ZipIR: Latent Pyramid Diffusion Transformer for High-Resolution Image Restoration** \\\n[[Website](https://arxiv.org/abs/2504.08591)]\n\n**Diffusion Once and Done: Degradation-Aware LoRA for Efficient All-in-One Image Restoration** \\\n[[Website](https://arxiv.org/abs/2508.03373)]\n\n**DiTVR: Zero-Shot Diffusion Transformer for Video Restoration** \\\n[[Website](https://arxiv.org/abs/2508.07811)]\n\n**Vivid-VR: Distilling Concepts from Text-to-Video Diffusion Transformer for Photorealistic Video Restoration** \\\n[[Website](https://arxiv.org/abs/2508.14483)]\n\n**OS-DiffVSR: Towards One-step Latent Diffusion Model for High-detailed Real-world Video Super-Resolution** \\\n[[Website](https://arxiv.org/abs/2509.16507)]\n\n**Boosting Fidelity for Pre-Trained-Diffusion-Based Low-Light Image Enhancement via Condition Refinement** \\\n[[Website](https://arxiv.org/abs/2510.17105)]\n\n**Noise is All You Need: Solving Linear Inverse Problems by Noise Combination Sampling with Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2510.23633)]\n\n**Enhancing Diffusion-based Restoration Models via Difficulty-Adaptive Reinforcement Learning with IQA Reward** \\\n[[Website](https://arxiv.org/abs/2511.01645)]\n\n**Integrating Reweighted Least Squares with Plug-and-Play Diffusion Priors for Noisy Image Restoration** \\\n[[Website](https://arxiv.org/abs/2511.06823)]\n\n**InstantViR: Real-Time Video Inverse Problem Solver with Distilled Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2511.14208)]\n\n**BokehFlow: Depth-Free Controllable Bokeh Rendering via Flow Matching** \\\n[[Website](https://arxiv.org/abs/2511.15066)]\n\n**UnfoldLDM: Deep Unfolding-based Blind Image Restoration with Latent Diffusion Priors** \\\n[[Website](https://arxiv.org/abs/2511.18152)]\n\n**CARD: Correlation Aware Restoration with Diffusion** \\\n[[Website](https://arxiv.org/abs/2512.05268)]\n\n**SURE Guided Posterior Sampling: Trajectory Correction for Diffusion-Based Inverse Problems** \\\n[[Website](https://arxiv.org/abs/2512.23232)]\n\n**Measurement-Consistent Langevin Corrector: A Remedy for Latent Diffusion Inverse Solvers** \\\n[[Website](https://arxiv.org/abs/2601.04791)]\n\n**Unifying Heterogeneous Degradations: Uncertainty-Aware Diffusion Bridge Model for All-in-One Image Restoration** \\\n[[Website](https://arxiv.org/abs/2601.21592)]\n\n**Zero-Shot Video Restoration and Enhancement with Assistance of Video Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2601.21922)]\n\n**LCUDiff: Latent Capacity Upgrade Diffusion for Faithful Human Body Restoration** \\\n[[Website](https://arxiv.org/abs/2602.04406)]\n\n## Colorization\n\n**LVCD: Reference-based Lineart Video Colorization with Diffusion Models** \\\n[[SIGGRAPH Asia 2024](https://arxiv.org/abs/2409.12960)]\n[[Project](https://luckyhzt.github.io/lvcd)]\n[[Code](https://github.com/luckyhzt/LVCD)]\n\n**Cobra: Efficient Line Art COlorization with BRoAder References** \\\n[[SIGGRAPH 2025](https://arxiv.org/abs/2504.12240)]\n[[Project](https://zhuang2002.github.io/Cobra/)]\n[[Code](https://github.com/Zhuang2002/Cobra)]\n\n**MagicColor: Multi-Instance Sketch Colorization** \\\n[[Website](https://arxiv.org/abs/2503.16948)]\n[[Project](https://yinhan-zhang.github.io/color/)]\n[[Code](https://github.com/YinHan-Zhang/MagicColor)]\n\n**ColorFlow: Retrieval-Augmented Image Sequence Colorization** \\\n[[Website](https://arxiv.org/abs/2412.11815)]\n[[Project](https://zhuang2002.github.io/ColorFlow/)]\n[[Code](https://github.com/TencentARC/ColorFlow)]\n\n**Control Color: Multimodal Diffusion-based Interactive Image Colorization** \\\n[[Website](https://arxiv.org/abs/2402.10855)]\n[[Project](https://zhexinliang.github.io/Control_Color/)]\n[[Code](https://github.com/ZhexinLiang/Control-Color)]\n\n**Multimodal Semantic-Aware Automatic Colorization with Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2404.16678)]\n[[Project](https://servuskk.github.io/ColorDiff-Image/)]\n[[Code](https://github.com/servuskk/ColorDiff-Image)]\n\n**MangaNinja: Line Art Colorization with Precise Reference Following** \\\n[[Website](https://arxiv.org/abs/2501.08332)]\n[[Project](https://johanan528.github.io/MangaNinjia/)]\n[[Code](https://github.com/ali-vilab/MangaNinjia)]\n\n**Multimodal Semantic-Aware Automatic Colorization with Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2404.16678)]\n[[Project](https://servuskk.github.io/ColorDiff-Image/)]\n[[Code](https://github.com/servuskk/ColorDiff-Image)]\n\n**VanGogh: A Unified Multimodal Diffusion-based Framework for Video Colorization** \\\n[[Website](https://arxiv.org/abs/2501.09499)]\n[[Project](https://becauseimbatman0.github.io/VanGogh)]\n[[Code](https://github.com/BecauseImBatman0/VanGogh)]\n\n**SketchColour: Channel Concat Guided DiT-based Sketch-to-Colour Pipeline for 2D Animation** \\\n[[Website](https://arxiv.org/abs/2507.01586)]\n[[Project](https://bconstantine.github.io/SketchColour/)]\n[[Code](https://github.com/bconstantine/SketchColour)]\n\n**L-CAD: Language-based Colorization with Any-level Descriptions using Diffusion Priors** \\\n[[Website](https://arxiv.org/abs/2305.15217)]\n[[Code](https://github.com/changzheng123/L-CAD)]\n\n**SSIMBaD: Sigma Scaling with SSIM-Guided Balanced Diffusion for AnimeFace Colorization** \\\n[[Website](https://arxiv.org/abs/2506.04283)]\n[[Code](https://github.com/Giventicket/SSIMBaD-Sigma-Scaling-with-SSIM-Guided-Balanced-Diffusion-for-AnimeFace-Colorization)]\n\n**Image Referenced Sketch Colorization Based on Animation Creation Workflow** \\\n[[Website](https://arxiv.org/abs/2502.19937)]\n[[Code](https://github.com/tellurion-kanata/colorizeDiffusion)]\n\n**ColorizeDiffusion: Adjustable Sketch Colorization with Reference Image and Text** \\\n[[Website](https://arxiv.org/abs/2401.01456)]\n[[Code](https://github.com/tellurion-kanata/colorizeDiffusion)]\n\n**ColorizeDiffusion v2: Enhancing Reference-based Sketch Colorization Through Separating Utilities** \\\n[[Website](https://arxiv.org/abs/2504.06895)]\n[[Code](https://github.com/tellurion-kanata/colorizeDiffusion)]\n\n**Leveraging the Powerful Attention of a Pre-trained Diffusion Model for Exemplar-based Image Colorization** \\\n[[Website](https://arxiv.org/abs/2505.15812)]\n[[Code](https://github.com/satoshi-kosugi/powerful-attention)]\n\n**Diffusing Colors: Image Colorization with Text Guided Diffusion** \\\n[[SIGGRAPH Asia 2023](https://arxiv.org/abs/2312.04145)]\n[[Project](https://pub.res.lightricks.com/diffusing-colors/)]\n\n**VanGogh: A Unified Multimodal Diffusion-based Framework for Video Colorization** \\\n[[Website](https://arxiv.org/abs/2501.09499)]\n[[Project](https://becauseimbatman0.github.io/VanGogh)]\n\n**Enhancing Diffusion Posterior Sampling for Inverse Problems by Integrating Crafted Measurements** \\\n[[Website](https://arxiv.org/abs/2411.09850)]\n\n**DiffColor: Toward High Fidelity Text-Guided Image Colorization with Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2308.01655)]\n\n**Consistent Video Colorization via Palette Guidance** \\\n[[Website](https://arxiv.org/abs/2501.19331)]\n\n**L-C4: Language-Based Video Colorization for Creative and Consistent Color** \\\n[[Website](https://arxiv.org/abs/2410.04972)]\n\n**LatentColorization: Latent Diffusion-Based Speaker Video Colorization** \\\n[[Website](https://arxiv.org/abs/2405.05707)]\n\n**Controllable Image Colorization with Instance-aware Texts and Masks** \\\n[[Website](https://arxiv.org/abs/2505.08705)]\n\n**AnimeColor: Reference-based Animation Colorization with Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2507.20158)]\n\n**MangaDiT: Reference-Guided Line Art Colorization with Hierarchical Attention in Diffusion Transformers** \\\n[[Website](https://arxiv.org/abs/2508.09709)]\n\n**Enhancing Reference-based Sketch Colorization via Separating Reference Representations** \\\n[[Website](https://arxiv.org/abs/2508.17620)]\n\n**Prompt-based Consistent Video Colorization** \\\n[[Website](https://arxiv.org/abs/2511.22330)]\n\n\n## Face Restoration\n\n**InterLCM: Low-Quality Images as Intermediate States of Latent Consistency Models for Effective Blind Face Restoration** \\\n[[ICLR 2025](https://openreview.net/forum?id=rUxr9Ll5FQ)]\n[[Website](https://arxiv.org/abs/2502.02215)]\n[[Project](https://sen-mao.github.io/InterLCM-Page/)]\n[[Code](https://github.com/sen-mao/InterLCM)]\n[[Demo](https://2lq8im394062.vicp.fun/)]\n\n**Self-Supervised Selective-Guided Diffusion Model for Old-Photo Face Restoration** \\\n[[Website](https://arxiv.org/abs/2510.12114)]\n[[Project](https://24wenjie-li.github.io/projects/SSDiff/)]\n[[Code](https://github.com/PRIS-CV/SSDiff)]\n\n**DiffBIR: Towards Blind Image Restoration with Generative Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2308.15070)]\n[[Project](https://0x3f3f3f3fun.github.io/projects/diffbir/)]\n[[Code](https://github.com/XPixelGroup/DiffBIR)]\n\n**OSDFace: One-Step Diffusion Model for Face Restoration** \\\n[[Website](https://arxiv.org/abs/2411.17163)]\n[[Project](https://jkwang28.github.io/OSDFace-web/)]\n[[Code](https://github.com/jkwang28/OSDFace)]\n\n**DR2: Diffusion-based Robust Degradation Remover for Blind Face Restoration** \\\n[[CVPR 2023](https://arxiv.org/abs/2303.06885)]\n[[Code](https://github.com/Kaldwin0106/DR2_Drgradation_Remover)]\n\n**PGDiff: Guiding Diffusion Models for Versatile Face Restoration via Partial Guidance** \\\n[[NeurIPS 2023](https://arxiv.org/abs/2309.10810)]\n[[Code](https://github.com/pq-yang/pgdiff)]\n\n**AT-DDPM: Restoring Faces degraded by Atmospheric Turbulence using Denoising Diffusion Probabilistic Models** \\\n[[WACV 2023](https://arxiv.org/abs/2208.11284)]\n[[Code](https://github.com/Nithin-GK/AT-DDPM)]\n\n**FLAIR: A Conditional Diffusion Framework with Applications to Face Video Restoration** \\\n[[WACV 2025](https://arxiv.org/abs/2311.15445)]\n[[Code](https://github.com/wustl-cig/FLAIR)]\n\n**HonestFace: Towards Honest Face Restoration with One-Step Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2505.18469)]\n[[Code](https://github.com/jkwang28/HonestFacee)]\n\n**DifFace: Blind Face Restoration with Diffused Error Contraction** \\\n[[Website](https://arxiv.org/abs/2312.15736)]\n[[Code](https://github.com/zsyOAOA/DifFace)]\n\n**AuthFace: Towards Authentic Blind Face Restoration with Face-oriented Generative Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2410.09864)]\n[[Code](https://github.com/EthanLiang99/AuthFace)]\n\n**Towards Real-World Blind Face Restoration with Generative Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2312.15736)]\n[[Code](https://github.com/chenxx89/BFRffusion)]\n\n**QuantFace: Low-Bit Post-Training Quantization for One-Step Diffusion Face Restoration** \\\n[[Website](https://arxiv.org/abs/2506.00820)]\n[[Code](https://github.com/jiatongli2024/QuantFace)]\n\n**Towards Unsupervised Blind Face Restoration using Diffusion Prior** \\\n[[Website](https://arxiv.org/abs/2410.04618)]\n[[Project](https://dt-bfr.github.io/)]\n\n**DynFaceRestore: Balancing Fidelity and Quality in Diffusion-Guided Blind Face Restoration with Dynamic Blur-Level Mapping and Guidance** \\\n[[ICCV 2025](https://arxiv.org/abs/2507.13797)]\n\n**InfoBFR: Real-World Blind Face Restoration via Information Bottleneck** \\\n[[Website](https://arxiv.org/abs/2501.15443)]\n\n**DiffBFR: Bootstrapping Diffusion Model Towards Blind Face Restoration** \\\n[[Website](https://arxiv.org/abs/2305.04517)]\n\n**CLR-Face: Conditional Latent Refinement for Blind Face Restoration Using Score-Based Diffusion Models** \\\n[[Website](https://arxiv.org/abs/2402.06106)]\n\n**DiffMAC: Diffusion Manifold Hallucination Correction for High Generalization Blind Face Restoration** \\\n[[Website](https://arxiv.org/abs/2403.10098)]\n\n**Gaussian is All You Need: A Unified Framework for Solving Inverse Problems via Diffusion Posterior Sampling** \\\n[[Website](https://arxiv.org/abs/2409.08906)]\n\n**Overcoming False Illusions in Real-World Face Restoration with Multi-Modal Guided Diffusion Model** \\\n[[Website](https://arxiv.org/abs/2410.04161)]\n\n**DR-BFR: Degradation Representation with Diffusion Models for Blind Face Restoration** \\\n[[Website](https://arxiv.org/abs/2411.10508)]\n\n**Face2Face: Label-driven Facial Retouching Restoration** \\\n[[Website](https://arxiv.org/abs/2404.14177)]\n\n**WaveFace: Authentic Face Restoration with Efficient Frequency Recovery** \\\n[[Website](https://arxiv.org/abs/2403.12760)]\n\n**DiffusionReward: Enhancing Blind Face Restoration through Reward Feedback Learning** \\\n[[Website](https://arxiv.org/abs/2505.17910)]\n\n**LAFR: Efficient Diffusion-based Blind Face Restoration via Latent Codebook Alignment Adapter** \\\n[[Website](https://arxiv.org/abs/2505.23462)]\n\n**Unlocking the Potential of Diffusion Priors in Blind Face Restoration** \\\n[[Website](https://arxiv.org/abs/2508.08556)]\n\n**BIR-Adapter: A Low-Complexity Diffusion Model Adapter for Blind Image Restoration** \\\n[[Website](https://arxiv.org/abs/2509.06904)]\n\n\n\n## Image Compression\n\n**Ultra Lowrate Image Compression with Semantic Residual Coding and Compression-aware Diffusion** \\\n[[ICML 2025](https://arxiv.org/abs/2505.08281)] \n[[Project](https://njuvision.github.io/ResULIC/)] \n[[Code](https://github.com/NJUVISION/ResULIC)] \n\n**DiT-IC: Aligned Diffusion Transformer for Efficient Image Compression** \\\n[[CVPR 2026](https://arxiv.org/abs/2603.13162)] \n[[Project](https://njuvision.github.io/DiT-IC/)] \n[[Code](https://github.com/Eric-qi/DiT-IC)] \n\n**OSCAR: One-Step Diffusion Codec Across Multiple Bit-rates** \\\n[[NeurIPS 2025](https://arxiv.org/abs/2505.16091)] \n[[Code](https://github.com/jp-guo/OSCAR)] \n\n**Towards Extreme Image Compression with Latent Feature Guidance and Diffusion Prior** \\\n[[IEE TCSVT 2024](https://arxiv.org/abs/2404.18820)] \n[[Code](https://github.com/huai-chang/DiffEIC)] \n\n**Taming Large Multimodal Agents for Ultra-low Bitrate Semantically Disentangled Image Compression** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.00399)] \n[[Code](https://github.com/yang-xidian/SEDIC)] \n\n**Using Powerful Prior Knowledge of Diffusion Model in Deep Unfolding Networks for Image Compressive Sensing** \\\n[[CVPR 2025](https://arxiv.org/abs/2503.08429)] \n","projects_url":"https://awesome.ecosyste.ms/api/v1/lists/wangkai930418%2Fawesome-diffusion-categorized/projects"}