Awesome-Generative-Image-Composition
A curated list of papers, code, and resources pertaining to generative image composition or object insertion.
https://github.com/bcmi/Awesome-Generative-Image-Composition
Last synced: 16 days ago
JSON representation
-
Survey
-
Papers
-
Evaluating Your Results
- [arXiv - vilab/AnyDoor) [[demo]](https://huggingface.co/spaces/xichenhku/AnyDoor-online)
- [arXiv
- [arXiv
- [arXiv - Studio/Paint-by-Example)
- [arXiv - hachnochi/cross-domain-compositing)
- [arXiv - Image-Composition) [[demo]](https://bcmi.sjtu.edu.cn/home/niuli/demo_image_composition/)
- [arXiv
- [arXiv - Image-Composition)
- [arXiv - vilab/AnyDoor) [[demo]](https://huggingface.co/spaces/xichenhku/AnyDoor-online)
-
Object-to-Object
- [arXiv - vilab/AnyDoor)
-
(Object+Text)-Guided
-
Training-based
- [arXiv
- [pdf - uofa/FreeCompose)
- [arXiv
- [paper
- [arXiv - wensong/insert-anything)
- [arXiv - aware-any)
- [arXiv
- [arXiv - tao/MotionCom)
- [arXiv - Image-Composition)
- [arXiv - World/UniCombine)
- [arXiv
- [arXiv - Image-Composition)
- [pdf - diffusion)
- [arXiv - Image-Composition) [[demo]](https://bcmi.sjtu.edu.cn/home/niuli/demo_image_composition/)
- [arXiv - Image-Composition)
- [arXiv - Studio/Paint-by-Example)
- [paper - research/affordance-insertion)
- [arXiv - tao/MotionCom)
- [arXiv - demo-page)
- [arXiv
- [arXiv - tao/MotionCom)
- [pdf - vilab/AnyDoor) [[demo]](https://huggingface.co/spaces/xichenhku/AnyDoor-online)
- [paper - research/affordance-insertion)
- [arXiv
-
Training-free
-
Object-Guided
- [arXiv - demo-page)
-
-
Foreground: image; Background: image
- [pdf - LU/TF-ICON)
-
Related Topics
-
Datasets
- SAM-FB - domain, single-ref): built upon SA-1B (SAM dataset). 3,160,403 images with 3,439 foreground categories.
- TF-ICON test benchmark - domain, single-ref): 332 samples. Each sample consists of a background image, a foreground image, a
- FOSCom - domain, single-ref): 640 background images from Internet. Each background image has a manually annotated bounding box and a foreground image from MSCOCO training set.
- MureCom - domain, multi-ref): 640 background images and 96 unique foreground objects from 32 categories.
- Subjects 200K - domain, double-ref): 200,000 paired images. Each pair has the same subject yet various scene contexts.
- DreamEditBench - domain, multi-ref): 220 background images and 30 unique foreground objects from 15 categories.
-
Evaluation Metrics
-
Leaderboard
-
Evaluating Your Results
-
-
Other Resources
-
Online Demo
Programming Languages
Categories
Sub Categories
Keywords
image-composition
6
object-insertion
4
diffusion-model
3
stable-diffusion
3
image-blending
2
image-harmonization
2
computer-vision
2
deep-learning
2
diffusion-models
2
image-editing
2
image-generation
2
image-manipulation
2
paint-by-example
2
pytorch
2
image-compositing
1
object-addition
1
object-compositing
1
text-to-image
1
object-placement
1
image-inversion
1
generative-ai
1
reflection-generation
1
shadow-generation
1