Research map / 3D vision / Human pose estimation
Coordinate-Based Texture Inpainting for Pose-Guided Human Image Generation
Artur Grigorev, Artem Sevastopolsky, Alexander Vakhitov, Victor Lempitsky
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 method instead of VUNet | Pose-guided human image generation on DeepFashion · DeepFashion validation | SSIM: 0.752 vs 0.835 | Improved |
| Use the proposed method instead of VUNet | Pose-guided human image generation on DeepFashion · DeepFashion validation | MS-SIM: 0.756 vs 0.802 | Improved |
| Use the proposed method instead of VUNet | Pose-guided human image generation on DeepFashion · DeepFashion validation | IS: 3.32 vs 2.93 | Hurt |
| Use the proposed method instead of VUNet | Pose-guided human image generation · User comparison against VUNet | preference rate: 59\% | Improved |
| Add NN loss (ablation) | Ours · User comparison against Ours without NN loss | preference rate: 79\% | Improved |
| Replace RGB-only inpainting with coordinate-based texture inpainting in the full system (ablation) | Pose-guided human image generation · User comparison against RGB-only inpainting | preference rate: 60\% | Improved |
| Use the full coordinate-based inpainting architecture instead of RGB-only inpainting (ablation) | Pose-guided human image generation · Validation sets | loss | Improved |
| Add NN loss (ablation) | Ours · Limited-pose-change user comparison against SwapNet | preference rate: 59\% | Improved |
| Use SwapNet instead of the submitted version of Ours | Pose-guided human image generation · Limited-pose-change user comparison | preference rate: 64\% | Hurt |
Closest papers by meaning
- Collecting The Puzzle Pieces: Disentangled Self-Driven Human Pose Transfer by Permuting Textures (ICCV 2023 · similarity 0.88)
- Neural Re-Rendering of Humans from a Single Image (ECCV 2020 · similarity 0.87)
- Pose-Guided Human Animation From a Single Image in the Wild (CVPR 2021 · similarity 0.87)
- SCULPT: Shape-Conditioned Unpaired Learning of Pose-dependent Clothed and Textured Human Meshes (CVPR 2024 · similarity 0.87)
- Disentangled Pose and Appearance Guidance for Multi-Pose Generation (CVPR 2025 · similarity 0.86)
- Pose Modulated Avatars from Video (ICLR 2024 · 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.
