Research map / Video / Video segmentation and tracking
EventGait: Towards Robust Gait Recognition with Event Streams
Senyan Xu, Shuai Chen, Chuanfu Shen, Kean Liu, Zhijing Sun, Chengzhi Cao, Xueyang Fu
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 dynamic motion stream to static stream (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 82.0 -> 92.8 | Improved |
| add static shape stream to dynamic stream (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 72.4 -> 92.8 | Improved |
| add CroSA structure alignment from DINOv2 teacher (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 87.4 -> 92.8 | Improved |
| use l2 instead of cosine alignment (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 89.1 -> 92.8 | Improved |
| raise alignment weight from 0.2 to 0.5 (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 92.8 -> 89.7 | Hurt |
| raise alignment weight from 0.05 to 0.2 (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 90.5 -> 92.8 | Improved |
| increase MoSE experts from 1 to 3 (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 88.4 -> 92.8 | Improved |
| increase MoSE experts from 3 to 4 (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 92.8 -> 92.7 | No clear effect |
| use event input instead of silhouettes with GaitBase (ablation) | EventGait · SUSTech1K-E | Overall Rank-1: 76.1 -> 63.1 | Hurt |
| use EventGait instead of GaitBase silhouettes in low light | EventGait vs GaitBase · SUSTech1K-E (low light) | Overall Rank-1: 41.5 -> 83.2 | Improved |
| use EventGait instead of silhouette methods at night in cross-domain | EventGait vs DeepGaitV2 · CCGR-Mini -> SUSTech1K (cross-domain) | NT Rank-1: 21.9 -> 51.4 | Improved |
| use EventGait instead of silhouette methods in cross-domain overall | EventGait vs DeepGaitV2 · CCGR-Mini -> SUSTech1K (cross-domain) | Overall Rank-1: 48.2 -> 45.5 | Mixed |
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.
- EVGait — EV-Gait: Event-Based Robust Gait Recognition Using Dynamic Vision Sensors (CVPR 2019) · better on 3 of 3 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
- EV-Gait: Event-Based Robust Gait Recognition Using Dynamic Vision Sensors (CVPR 2019 · similarity 0.93)
- Event-Stream Representation for Human Gaits Identification Using Deep Neural Networks (TPAMI 2021 · similarity 0.93)
- Gait Recognition in the Wild: A Large-Scale Benchmark and NAS-Based Baseline (TPAMI 2025 · similarity 0.89)
- Learning Gait Representation From Massive Unlabelled Walking Videos: A Benchmark (TPAMI 2023 · similarity 0.88)
- Learning Visual Prompt for Gait Recognition (CVPR 2024 · similarity 0.87)
- BiggerGait: Unlocking Gait Recognition with Layer-wise Representations from Large Vision Models (NeurIPS 2025 · similarity 0.86)
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.
