Research map / Video / Action recognition
Gait Recognition via Effective Global-Local Feature Representation and Local Temporal Aggregation
Beibei Lin, Shunli Zhang, Xin Yu
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 local branch to global branch in GLConv (ablation) | GLFE module · CASIA-B LT(74) NM | Rank-1 accuracy: 97.1 -> 97.4 | Improved |
| add global branch to local branch in GLConv (ablation) | GLFE module · CASIA-B LT(74) NM | Rank-1 accuracy: 96.2 -> 97.4 | Improved |
| add local branch to global branch in GLConv (ablation) | GLFE module · CASIA-B LT(74) CL | Rank-1 accuracy: 81.9 -> 83.6 | Improved |
| add global branch to local branch in GLConv (ablation) | GLFE module · CASIA-B LT(74) CL | Rank-1 accuracy: 80.4 -> 83.6 | Improved |
| vary number of local partitions N from 2 to 8 (ablation) | GLFE module · CASIA-B LT(74) NM | Rank-1 accuracy: 97.3 -> 97.4 | No clear effect |
| replace Max pooling with GeM pooling in spatial feature mapping (ablation) | feature mapping · CASIA-B LT(74) mean over NM/BG/CL | Rank-1 accuracy: 91.3 -> 91.8 | Improved |
| replace Avg pooling with GeM pooling (ablation) | feature mapping · CASIA-B LT(74) mean over NM/BG/CL | Rank-1 accuracy: 88.5 -> 91.8 | Improved |
| replace weighted Max+Avg with GeM (ablation) | feature mapping · CASIA-B LT(74) CL | Rank-1 accuracy: 82.8 -> 83.6 | Improved |
| replace first spatial pooling with LTA (ablation) | downsampling · CASIA-B LT(74) NM | Rank-1 accuracy: 96.3 -> 97.4 | Improved |
| replace first spatial pooling with LTA (ablation) | downsampling · CASIA-B LT(74) CL | Rank-1 accuracy: 77.7 -> 83.6 | Improved |
| replace both spatial pooling layers with LTA (ablation) | downsampling · CASIA-B LT(74) NM | Rank-1 accuracy: 97.4 -> 94.4 | Hurt |
| place LTA at second instead of first pooling position (ablation) | downsampling · CASIA-B LT(74) NM | Rank-1 accuracy: 97.4 -> 97.0 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GaitPart — GaitPart: Temporal Part-Based Model for Gait Recognition (CVPR 2020) · better on 4 of 4 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
- Hierarchical Spatio-Temporal Representation Learning for Gait Recognition (ICCV 2023 · similarity 0.91)
- Learning Gait Representation From Massive Unlabelled Walking Videos: A Benchmark (TPAMI 2023 · similarity 0.90)
- Deep Gait Recognition: A Survey (TPAMI 2022 · similarity 0.90)
- Context-Sensitive Temporal Feature Learning for Gait Recognition (ICCV 2021 · similarity 0.90)
- Learning Visual Prompt for Gait Recognition (CVPR 2024 · similarity 0.89)
- BiggerGait: Unlocking Gait Recognition with Layer-wise Representations from Large Vision Models (NeurIPS 2025 · similarity 0.88)
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.
