Research map / Image generation / GANs
XingGAN for Person Image Generation
Hao Tang, Song Bai, Li Zhang, Philip H.S. Torr, Nicu Sebe
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add the Appearance-guided Shape branch to the Shape-guided Appearance branch (ablation) | XingGAN SA · Market-1501 | SSIM: 0.239 -> 0.310 | Improved |
| use AS branch instead of SA branch alone (ablation) | XingGAN · Market-1501 | SSIM: 0.239 -> 0.286 | Improved |
| add the SA branch to the AS branch (dual-branch) (ablation) | XingGAN AS · Market-1501 | Mask-SSIM: 0.798 -> 0.807 | Improved |
| add co-attention fusion (CAF) to SA+AS (ablation) | XingGAN · Market-1501 | SSIM: 0.310 -> 0.313 | Improved |
| add co-attention fusion (CAF) to SA+AS (ablation) | XingGAN · Market-1501 | Mask-SSIM: 0.807 -> 0.816 | Improved |
| add co-attention fusion (CAF) to SA+AS (ablation) | XingGAN · Market-1501 | IS: 3.558 -> 3.506 | Hurt |
| increase Xing blocks from 1 to 9 (ablation) | Xing Generator · Market-1501 | Mask-SSIM: 0.812 -> 0.816 | Improved |
| increase Xing blocks from 9 to 13 (ablation) | Xing Generator · Market-1501 | SSIM: 0.313 -> 0.257 | Hurt |
| replace Resnet blocks with Xing blocks (5 blocks vs 13 Resnet blocks) (ablation) | Resnet Generator · Market-1501 | SSIM: 0.300 -> 0.313 | Improved |
| replace PATN blocks with Xing blocks (5 blocks vs 13 PATN blocks) (ablation) | PATN Generator · Market-1501 | Mask-SSIM: 0.808 -> 0.812 | Improved |
| use XingGAN instead of Pose-Transfer | Pose-Transfer · Market-1501 | SSIM: 0.311 -> 0.313 | Improved |
| use XingGAN instead of Pose-Transfer | Pose-Transfer · Market-1501 | PCKh: 0.94 -> 0.93 | Hurt |
Closest papers by meaning
- Enhanced Multi-Scale Cross-Attention for Person Image Generation (TPAMI 2025 · similarity 0.92)
- Appearance and Pose-Conditioned Human Image Generation Using Deformable GANs (TPAMI 2019 · similarity 0.89)
- BodyGAN: General-Purpose Controllable Neural Human Body Generation (CVPR 2022 · similarity 0.88)
- InsetGAN for Full-Body Image Generation (CVPR 2022 · similarity 0.87)
- Deformable GANs for Pose-Based Human Image Generation (CVPR 2018 · similarity 0.87)
- RealisHuman: A Two-Stage Approach for Refining Malformed Human Parts in Generated Images (AAAI 2025 · similarity 0.85)
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.
