AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

F-Bench: Rethinking Human Preference Evaluation Metrics for Benchmarking Face Generation, Customization, and Restoration

Lu Liu, Huiyu Duan, Qiang Hu, Liu Yang, Chunlei Cai, Tianxiao Ye, Huayu Liu, Xiaoyun Zhang, Guangtao Zhai

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Fairness and bias in generative models · 11 citations (OpenAlex)

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
add the face encoder (ArcFace) (ablation)F-Eval · FaceQ-Res real-world ID FidelitySRCC: 0.603 -> 0.796Improved
add the projector (ablation)F-Eval · FaceQ-Res real-world ID FidelitySRCC: 0.478 -> 0.796Improved
use four dimension-wise LoRA experts instead of single LoRA (ablation)F-Eval · FaceQ-Res real-world ID FidelitySRCC: 0.671 -> 0.796Improved
use four dimension-wise LoRA experts instead of three task-wise LoRA (ablation)F-Eval · FaceQ-Res real-world ID FidelitySRCC: 0.638 -> 0.796Improved
use four dimension-wise LoRA experts instead of single LoRA (ablation)F-Eval · FaceQ-Gen QualitySRCC: 0.793 -> 0.849Improved
add the face encoder (ablation)F-Eval · FaceQ-Gen QualitySRCC: 0.815 -> 0.849Improved
add the projector (ablation)F-Eval · FaceQ-Gen QualitySRCC: 0.808 -> 0.849Improved
add face encoder on face generation correspondence (ablation)F-Eval · FaceQ-Gen CorrespondenceSRCC: 0.835 -> 0.847No clear effect
use dimension LoRA instead of single LoRA on customization quality (ablation)F-Eval · FaceQ-Cus QualitySRCC: 0.941 -> 0.946No clear effect
fine-tune MINT-IQA on FaceQMINT-IQA · FaceQ-Gen QualitySRCC: 0.7002 -> 0.8312Improved
replace best baseline with F-Eval on customization ID fidelityPromptIQA (fine-tuned) · FaceQ-Cus ID FidelitySRCC: 0.8693 -> 0.9188Improved
replace best baseline with F-Eval on real-world restoration qualityMANIQA (fine-tuned) · FaceQ-Res real-world QualitySRCC: 0.8356 -> 0.8448Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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