Research map / Video / Video segmentation and tracking
STA: Spatial-Temporal Attention for Large-Scale Video-Based Person Re-Identification
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 hard-batch triplet loss to softmax training (ablation) | ResNet50 video Re-ID baseline · MARS | mAP: 64.0 -> 74.0 | Helped |
| Add hard-batch triplet loss to softmax training (ablation) | ResNet50 video Re-ID baseline · DukeMTMC-VideoReID | mAP: 76.8 -> 89.7 | Helped |
| Add average pooling across frames (ablation) | Baseline + triplet loss · MARS | mAP: 74.0 -> 75.0 | Helped |
| Replace average pooling with spatial-temporal attention (ablation) | Baseline + triplet loss · MARS | mAP: 75.0 -> 78.0 | Helped |
| Replace average pooling with spatial-temporal attention (ablation) | Baseline + triplet loss · DukeMTMC-VideoReID | mAP: 91.0 -> 92.7 | Helped |
| Add feature fusion combining discriminative and global information (ablation) | STA with triplet loss · MARS | mAP: 78.0 -> 79.1 | Helped |
| Add feature fusion combining discriminative and global information (ablation) | STA with triplet loss · DukeMTMC-VideoReID | mAP: 92.7 -> 93.9 | Helped |
| Add inter-frame regularization (ablation) | STA with feature fusion · MARS | mAP: 79.1 -> 80.8 | Helped |
| Add inter-frame regularization (ablation) | STA with feature fusion · DukeMTMC-VideoReID | mAP: 93.9 -> 94.9 | Helped |
| Use sequence length N=2 instead of N=4 (ablation) | STA · MARS | mAP: 80.8 -> 75.7 | Hurt |
| Use sequence length N=6 or N=8 instead of N=4 (ablation) | STA · MARS | mAP: 80.8 -> 81.0; 80.8 -> 81.2 | Helped |
| Use sequence length N=2 instead of N=4 (ablation) | STA · DukeMTMC-VideoReID | mAP: 94.9 -> 89.0 | Hurt |
Closest papers by meaning
- Diversity Regularized Spatiotemporal Attention for Video-Based Person Re-Identification (CVPR 2018 · similarity 0.91)
- Spatial-Temporal Graph Convolutional Network for Video-Based Person Re-Identification (CVPR 2020 · similarity 0.89)
- Relation-Guided Spatial Attention and Temporal Refinement for Video-Based Person Re-Identification (AAAI 2020 · similarity 0.88)
- Spatial-Temporal Person Re-Identification (AAAI 2019 · similarity 0.88)
- CLIMB-ReID: A Hybrid CLIP-Mamba Framework for Person Re-Identification (AAAI 2025 · similarity 0.87)
- Temporal Correlation Vision Transformer for Video Person Re-Identification (AAAI 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.