Research map / 3D vision / Face and shape
Portrait Eyeglasses and Shadow Removal by Leveraging 3D Synthetic Data
Junfeng Lyu, Zhibo Wang, Feng Xu
Check 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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use the proposed mask-guided, multi-step removal method | portrait eyeglasses and shadow removal · FFHQ | MOS: 3.17 -> 4.43 | Improved |
| Use the proposed method rather than HiSD | portrait eyeglasses and shadow removal · FFHQ | FID: 26.74 -> 26.89 | No clear effect |
| Remove the domain adaptation network (ablation) | proposed method · FFHQ | FID: 26.89 -> 27.45 | Hurt |
| Predict both masks together instead of predicting eyeglasses first to guide shadow prediction (ablation) | mask prediction stage · FFHQ | FID: 26.89 -> 27.18 | Hurt |
| Predict shadow masks first and use them to guide eyeglasses mask prediction (ablation) | mask prediction stage · FFHQ | FID: 26.89 -> 27.30 | Hurt |
| Add shadow-mask guidance to item removal (ablation) | proposed method · FFHQ | FID: 33.89 -> 26.89 | Improved |
| Add eyeglasses-mask guidance to item removal (ablation) | proposed method · FFHQ | FID: 42.80 -> 26.89 | Improved |
| Use separate sequential networks rather than one network for both removals (ablation) | item removal stage · FFHQ | FID: 28.66 -> 26.89 | Improved |
| Remove shadows before eyeglasses rather than eyeglasses first (ablation) | item removal stage · FFHQ | FID: 29.58 -> 26.89 | Improved |
| Add the mask operation before the De-Glass Network (ablation) | item removal stage · FFHQ | FID: 27.57 -> 26.89 | Improved |
Closest papers by meaning
- Unwarping the Lens: A Physics-Grounded Approach to Video Glasses Removal (ECCV 2026 · similarity 0.87)
- Polarization Guided Mask-Free Shadow Removal (AAAI 2025 · similarity 0.83)
- PHR-DIFF: Portrait Highlights Removal via Patch-aware Diffusion Model (AAAI 2025 · similarity 0.83)
- OmniSR: Shadow Removal Under Direct and Indirect Lighting (AAAI 2025 · similarity 0.82)
- DeS3: Adaptive Attention-Driven Self and Soft Shadow Removal Using ViT Similarity (AAAI 2024 · similarity 0.81)
- MetaShadow: Object-Centered Shadow Detection, Removal, and Synthesis (CVPR 2025 · similarity 0.81)
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.
