{"id":79726,"url":"https://github.com/andrewwango/awesome-unsupervised-imaging","name":"awesome-unsupervised-imaging","description":"Resources for solving imaging inverse problems using deep learning without ground truth","projects_count":36,"last_synced_at":"2026-09-24T06:00:24.875Z","repository":{"id":263606838,"uuid":"881341869","full_name":"Andrewwango/awesome-unsupervised-imaging","owner":"Andrewwango","description":"Resources for solving imaging inverse problems using deep learning without ground truth","archived":false,"fork":false,"pushed_at":"2024-11-20T22:17:47.000Z","size":6,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2026-09-04T08:06:28.936Z","etag":null,"topics":["awesome-list","computational-imaging","deep-learning","inverse-problems","medical-imaging","unsupervised-learning"],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"bsd-3-clause","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Andrewwango.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","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}},"created_at":"2024-10-31T11:38:20.000Z","updated_at":"2024-12-15T13:30:04.000Z","dependencies_parsed_at":"2024-11-26T21:02:36.653Z","dependency_job_id":null,"html_url":"https://github.com/Andrewwango/awesome-unsupervised-imaging","commit_stats":null,"previous_names":["andrewwango/awesome-unsupervised-imaging"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/Andrewwango/awesome-unsupervised-imaging","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Andrewwango%2Fawesome-unsupervised-imaging","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Andrewwango%2Fawesome-unsupervised-imaging/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Andrewwango%2Fawesome-unsupervised-imaging/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Andrewwango%2Fawesome-unsupervised-imaging/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Andrewwango","download_url":"https://codeload.github.com/Andrewwango/awesome-unsupervised-imaging/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Andrewwango%2Fawesome-unsupervised-imaging/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":341189360,"owners_count":37639594,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-08-22T15:14:58.755Z","status":"online","status_checked_at":"2026-09-24T02:00:06.325Z","response_time":86,"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":"2024-11-22T10:15:44.821Z","updated_at":"2026-09-24T06:00:24.884Z","primary_language":null,"list_of_lists":false,"displayable":true,"categories":["Frameworks","Overviews","Metrics","Foundations"],"sub_categories":["Generative models","Equivariant imaging","Multi-operator algorithms","Group theory"],"readme":"# Awesome Unsupervised Imaging\nA collection of papers, tutorials and software for solving imaging inverse problems using deep learning **without ground truth**.\n\nThe focus is learning frameworks (e.g. losses, training methods) and applications, but not specific architectures.\n\nContents\n1. [Overviews](#overviews)\n2. [Frameworks](#frameworks)\n3. [Metrics](#metrics)\n4. [Applications](#applications)\n5. [Foundations](#foundations)\n\n## Overviews\n\nTachella J. | Davies M. | [Tutorial: self-supervised learning for imaging](https://tachella.github.io/blog/selfsuptutorial/), 2024.\n- A good overview of current methods and theory of self-supervised learning from incomplete data.\n\nOngie G. et al. | [Deep learning techniques for inverse problems in imaging](https://arxiv.org/abs/2005.06001), 2020.\n- A 2020 taxonomy of supervised and unsupervised methods.\n\nChen D. | [Imaging with equivariant deep learning: from unrolled network design to fully unsupervised learning](https://ieeexplore.ieee.org/document/10004796), section _Computational imaging, equivariance, and deep learning_, 2023.\n- A brief timeline of classical methods to deep learning.\n\n## Frameworks\n\n### Equivariant imaging\n\nChen D. et al. | [Equivariant Imaging: Learning Beyond the Range Space](https://arxiv.org/abs/2103.14756), 2021.\n\nChen D. et al. | [Robust Equivariant Imaging: a fully unsupervised framework for learning to image from noisy and partial measurements](http://arxiv.org/abs/2111.12855), 2022.\n\nTachella J. | [Sensing Theorems for Unsupervised Learning in Linear Inverse Problems](http://arxiv.org/abs/2203.12513), 2023.\n\nWang A. et al. | [Perspective-Equivariance for Unsupervised Imaging with Camera Geometry](https://arxiv.org/abs/2403.09327), 2024.\n\nWang A. et al. | [Fully Unsupervised Dynamic MRI Reconstruction via Diffeo-Temporal Equivariance](https://arxiv.org/abs/2410.08646), 2024.\n\nScanvic J. et al. | [Self-Supervised Learning for Image Super-Resolution and Deblurring](https://arxiv.org/abs/2312.11232), 2024.\n\nSechaud V. et al. | [Equivariance-based self-supervised learning for audio signal recovery from clipped measurements](https://arxiv.org/abs/2409.15283), 2024.\n\n- Code tutorial: [Self-supervised learning with Equivariant Imaging for MRI | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_equivariant_imaging.html)\n- Code tutorial: [Image transformations for Equivariant Imaging | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_ei_transforms.html)\n\n### Multi-operator algorithms\n\n- Yaman + example\n- Millard \u0026 Chiew, Noise2Self, Noise2Inverse\n- Noise2Inverse, Noise2Void, Noise2Self\n- Artifact2Artifact, Phase2Phase + example\n- AmbientDiffusion - refer to tutorial\n- Tachella J. et al. | [Unsupervised Learning From Incomplete Measurements for Inverse Problems](https://arxiv.org/abs/2201.12151)\n\n- Code tutorial: [Self-supervised learning with measurement splitting | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_splitting_loss.html)\n- Code tutorial: [Self-supervised MRI reconstruction with Artifact2Artifact | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_artifact2artifact.html)\n- Code tutorial: [Self-supervised learning from incomplete measurements of multiple operators | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_multioperator_imaging.html)\n\n#### Denoising algorithms\n\nLe Montagner et al. | [An Unbiased Risk Estimator for Image Denoising in the Presence of Mixed Poisson-Gaussian Noise](https://ieeexplore.ieee.org/abstract/document/6714502)\n\nTachella J. et al. | [UNSURE: Unknown Noise level Stein's Unbiased Risk Estimator](https://arxiv.org/abs/2409.01985)\n\nEldar Y. | [Generalized SURE for Exponential Families: Applications to Regularization](https://ieeexplore.ieee.org/document/4663926)\n\nLehtinen J. et al. | [Noise2Noise: Learning Image Restoration without Clean Data](https://arxiv.org/abs/1803.04189)\n\nHuang T. et al. | [Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Images](https://arxiv.org/abs/2101.02824), 2021.\n\nPang T. et al. | [Recorrupted-to-Recorrupted: Unsupervised Deep Learning for Image Denoising](https://openaccess.thecvf.com/content/CVPR2021/papers/Pang_Recorrupted-to-Recorrupted_Unsupervised_Deep_Learning_for_Image_Denoising_CVPR_2021_paper.pdf)\n\n- Code tutorial: [Self-supervised denoising with the SURE loss | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_sure_denoising.html)\n- Code tutorial: [Self-supervised denoising with the UNSURE loss | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_unsure.html)\n- Code tutorial: [Self-supervised denoising with the Neighbor2Neighbor loss | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/self-supervised-learning/demo_n2n_denoising.html)\n\n### Generative models\n\n- Deep Image Prior\n- CoIL, Ultra-NeRF\n\n#### Generative adversarial models\n\nBora A. et al. | [AmbientGAN: Generative models from lossy measurements](https://openreview.net/forum?id=Hy7fDog0b)\n\nPajot A. et al. | [Unsupervised Adversarial Image Reconstruction](https://openreview.net/forum?id=BJg4Z3RqF7)\n\nLi B. et al. | [Progressive dual-domain-transfer cycleGAN for unsupervised MRI reconstruction](https://www.sciencedirect.com/science/article/abs/pii/S0925231223010573)\n\n#### Diffusion models\n\nDaras G. et al. | [Ambient Diffusion: Learning Clean Distributions from Corrupted Data](https://arxiv.org/abs/2305.19256)\n\nKawar B. et al. | [GSURE-Based Diffusion Model Training with Corrupted Data](https://arxiv.org/abs/2305.13128)\n\n- Code tutorial: [Imaging inverse problems with adversarial networks | DeepInverse](https://deepinv.github.io/deepinv/auto_examples/adversarial-learning/demo_gan_imaging.html)\n\nAside: Deep model-based approaches (DMBA)\n\n- Concise overview in [CoIL paper](https://arxiv.org/pdf/2102.05181 \"https://arxiv.org/pdf/2102.05181\") section 2\n- [Kamilov 2022 DMBA Overview](https://arxiv.org/abs/2203.17061 \"https://arxiv.org/abs/2203.17061\")\n- Inference-time optimisation procedures: (note these are not directly competitive with above frameworks, as these aren’t used to train the model - they assume a separately trained model, either sup or unsupervised)\n    - PnP, RED, RARE\n    - DRP\n    - SelfDEQ?\n\n- Architectures:\n    - Deep unrolling/unfolding\n    - Deep cascade\n\n## Metrics\n\n- [Perceptual-distortion trade-off](https://openaccess.thecvf.com/content_cvpr_2018/papers/Blau_The_Perception-Distortion_Tradeoff_CVPR_2018_paper.pdf \"https://openaccess.thecvf.com/content_cvpr_2018/papers/Blau_The_Perception-Distortion_Tradeoff_CVPR_2018_paper.pdf\") [video](https://www.youtube.com/watch?v=3Oetx8DCUCA \"https://www.youtube.com/watch?v=3Oetx8DCUCA\")\n\n## Applications\n\n### Magnetic Resonance Imaging\n\n- Find a good paper introducing the physics etc.\n- Good supervised/unsupervised/classical overview tutorial\n\n### Computed Tomography\n\n### Remote sensingg\n\n- Pan sharpening: structural measurement consistency, PanGan, QNR\n\n## Foundations\n\n### Linear algebra\n\n### Group theory\n\n- [Introduction to Groups, Representations, Equivariance and Geometric Deep Learning](https://geometricdeeplearning.com/slides/Cambridge_1_Introduction_to_Groups_and_Representations.pdf \"https://geometricdeeplearning.com/slides/Cambridge_1_Introduction_to_Groups_and_Representations.pdf\")\n- Chua lecture notes: [Group Theory](https://dec41.user.srcf.net/notes/IA_M/groups.pdf \"https://dec41.user.srcf.net/notes/IA_M/groups.pdf\"), [Group representation theory](https://dec41.user.srcf.net/notes/II_L/representation_theory.pdf \"https://dec41.user.srcf.net/notes/II_L/representation_theory.pdf\")","projects_url":"https://awesome.ecosyste.ms/api/v1/lists/andrewwango%2Fawesome-unsupervised-imaging/projects"}