Research map / Image generation / Image restoration
Gait Recognition from a Single Image using a Phase-Aware Gait Cycle Reconstruction Network
Chi Xu, Yasushi Makihara, Xiang Li, Yasushi Yagi, Jianfeng Lu
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 PA-GCR gait cycle reconstruction (ablation) | GaitSet · OU-MVLP | Rank-1 identification rate (%): 34.4 -> 80.3 | Improved |
| add unified (joint) training of PA-GCR and GaitSet (ablation) | PA-GCRNet · OU-MVLP | Rank-1 identification rate (%): 50.5 -> 80.3 | Improved |
| add phase estimator, feature transformer and similarity loss (ablation) | PA-GCRNet · OU-MVLP | Rank-1 identification rate (%): 77.4 -> 80.3 | Improved |
| use ground-truth phase instead of estimated phase (ablation) | PA-GCRNet · OU-MVLP | Rank-1 identification rate (%): 80.3 -> 80.6 | No clear effect |
| feed ground-truth gait cycle instead of reconstruction (ablation) | GaitSet · OU-MVLP | Rank-1 identification rate (%): 80.3 -> 97.7 | Improved |
| fine-tune PA-GCR from OU-MVLP pretrained model instead of training from scratch | PA-GCRNet · CASIA-B | Rank-1 identification rate (%): 39.4 -> 74.7 | Improved |
| fine-tune PA-GCR from OU-MVLP pretrained model instead of training from scratch | PA-GCRNet · OUTD-D | Rank-1 identification rate (%): 73 -> 91 | Improved |
| cross-dataset testing of OU-MVLP model instead of training from scratch | PA-GCRNet · OUTD-D | Rank-1 identification rate (%): 73 -> 75 | Improved |
| cross-dataset testing of OU-MVLP model instead of training from scratch | PA-GCRNet · CASIA-B | Rank-1 identification rate (%): 39.4 -> 74.7 | Improved |
| train PA-GCRNet from scratch on only 24 subjects vs GaitSet | PA-GCRNet vs GaitSet · CASIA-B | Rank-1 identification rate (%): 33.3 -> 39.4 | Improved |
Closest papers by meaning
- Learning Gait Representation From Massive Unlabelled Walking Videos: A Benchmark (TPAMI 2023 · similarity 0.87)
- Gait Recognition in the Wild: A Large-Scale Benchmark and NAS-Based Baseline (TPAMI 2025 · similarity 0.86)
- On Learning Disentangled Representations for Gait Recognition (TPAMI 2020 · similarity 0.86)
- Cut out the Middleman: Revisiting Pose-based Gait Recognition (ECCV 2024 · similarity 0.86)
- GaitCycFormer: Leveraging Gait Cycles and Transformers for Gait Emotion Recognition. (AAAI 2025 · similarity 0.85)
- Learning Visual Prompt for Gait Recognition (CVPR 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.
