{"id":13771651,"url":"https://github.com/bcmi/Composite-Image-Evaluation","last_synced_at":"2025-05-11T04:30:48.881Z","repository":{"id":188487237,"uuid":"678820850","full_name":"bcmi/Composite-Image-Evaluation","owner":"bcmi","description":null,"archived":false,"fork":false,"pushed_at":"2024-02-14T14:09:01.000Z","size":6,"stargazers_count":20,"open_issues_count":0,"forks_count":0,"subscribers_count":4,"default_branch":"main","last_synced_at":"2024-11-17T07:34:24.144Z","etag":null,"topics":["image-composition"],"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/bcmi.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}},"created_at":"2023-08-15T13:08:57.000Z","updated_at":"2024-10-24T09:03:48.000Z","dependencies_parsed_at":"2023-08-15T15:16:16.178Z","dependency_job_id":"4c097870-38ab-47f2-9d85-acfb243f8eaa","html_url":"https://github.com/bcmi/Composite-Image-Evaluation","commit_stats":null,"previous_names":["bcmi/composite-image-evaluation"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bcmi%2FComposite-Image-Evaluation","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bcmi%2FComposite-Image-Evaluation/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bcmi%2FComposite-Image-Evaluation/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bcmi%2FComposite-Image-Evaluation/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/bcmi","download_url":"https://codeload.github.com/bcmi/Composite-Image-Evaluation/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253518941,"owners_count":21921074,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","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"}},"keywords":["image-composition"],"created_at":"2024-08-03T17:00:53.947Z","updated_at":"2025-05-11T04:30:48.602Z","avatar_url":"https://github.com/bcmi.png","language":null,"funding_links":[],"categories":["Evaluation","Evaluation Metrics"],"sub_categories":[],"readme":"# Composite-Image-Evaluation\n\nHere are some possible evaluation metrics to evaluate the quality of composite images from different aspects. \n\n+ Evaluate whether the foreground is harmonious with background.\n\n  + Harmony score: use [illumination encoder](https://github.com/bcmi/BargainNet-Image-Harmonization) to extract the illumination codes from foreground and background, and measure their similarity.\n \n  + Inharmony hit: use [inharmonious region localization model](https://github.com/bcmi/MadisNet-Inharmonious-Region-Localization) to detect the inharmonious region, and calculate the overlap (e.g., IoU) between detected region and foreground region. \n\n+ Evaluate whether the foreground object placement is reasonable.\n\n  + OPA score: use [object placement assessment model](https://github.com/bcmi/Object-Placement-Assessment-Dataset-OPA) to predict the accuracy of object placement. \n\n+ Evaluate whether the foreground is compatible with background in terms of geometry and semantics.\n\n  + FOS score: use [foreground object search model](https://github.com/bcmi/Foreground-Object-Search-Dataset-FOSD) to calculate the compatibility score between foreground and background in terms of geometry and semantics. \n\n+ Evaluate the fidelity of foreground, i.e., whether the synthesized foreground is similar to the input foreground.\n\n  + Clip score: use [CLIP](https://github.com/openai/CLIP) to extract the embeddings from the input foreground image and the generated foreground patch, and measure their similarity.\n \n  + Dino score: use [DINO](https://github.com/facebookresearch/dino) to measure the average cosine similarity between the input and generated foreground.  \n\u003c!--\nSpecifically, we crop out foreground patch from generated composite and mask the non-object region to compute CLIP score with the masked input foreground, in which we estimate the mask of foreground object using [SAM](https://github.com/facebookresearch/segment-anything). We employ pretrained [clip-ViT-B-32](https://huggingface.co/openai/clip-vit-base-patch32) as image encoder following [Paint-by-Example](https://github.com/Fantasy-Studio/Paint-by-Example/tree/main).\n--\u003e\n\n+ Evaluate the over quality of foreground or the whole composite image.\n\n  + FID: use pretrained image encoder (*e.g.*, InceptionNet, CLIP) to extract the embeddings from real images and generated images, and measure their [Fréchet Inception Distance](https://github.com/mseitzer/pytorch-fid).\n  + QS: use [quality score](https://github.com/cientgu/GIQA) to measure the quality of each single generated image, and compute average score.    \n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbcmi%2FComposite-Image-Evaluation","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbcmi%2FComposite-Image-Evaluation","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbcmi%2FComposite-Image-Evaluation/lists"}