AI Paper Atlas

Research map / Image generation / Editing and style transfer

FaceComposer: A Unified Model for Versatile Facial Content Creation

Jiayu Wang, Kang Zhao, Yifeng Ma, Shiwei Zhang, Yingya Zhang, Yujun Shen, Deli Zhao, Jingren Zhou

NeurIPS 2023 · Image generation › Editing and style transfer · approach: Diffusion face generation and editing

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
remove image data (train on videos only) (ablation)FaceComposer · HDTF face animationM-LMD: 2.25 -> 2.46Hurt
remove video data (train on images only) (ablation)FaceComposer · HDTF face animationF-LMD: 1.84 -> 1.94Hurt
remove video data (train on images only) (ablation)FaceComposer · HDTF face animationCPBD: 0.27 -> 0.29No clear effect
remove image data (ablation)FaceComposer · HDTF face animationSSIM: 0.78 -> 0.78No clear effect
drop ID condition (ablation)FaceComposer · Face generationFID: 11.34 -> 11.40No clear effect
drop ID and Sketch conditions (ablation)FaceComposer · Face generationFID: 11.34 -> 11.27No clear effect
drop ID and Sketch conditions (ablation)FaceComposer · Face animationF-LMD: 1.84 -> 1.88Hurt
reduce data to 10 h video and 4.5W images (ablation)FaceComposer- · Face generationFID: 11.34 -> 11.60Hurt
reduce data to 10 h video and 4.5W images (ablation)FaceComposer- · Face animationM-LMD: 2.25 -> 2.37Hurt
FaceComposer instead of LAFITEface generation · Multi-Modal CelebA-HQR-precision (%): 81.94 -> 86.63Improved
FaceComposer instead of StyleClipface editing · Edited imagesIDS: 0.75 -> 0.94Improved
FaceComposer instead of Wav2Lipface animation · HDTFSyncconf: 5.27 -> 4.27Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy