Research map / Video / Video segmentation and tracking
BiCnet-TKS: Learning Efficient Spatial-Temporal Representation for Video Person Re-Identification
Check 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 Cross-Scale Paths to the two-branch network (ablation) | BiCnet · MARS | mAP: 84.3 -> 85.0 | Helped |
| Add diverse attention modules without diversity regularization (ablation) | TB+CSP · MARS | mAP: 85.0 -> 85.2 | No clear effect |
| Add a diversity loss to guide complementary attention maps (ablation) | TB+CSP+AO · MARS | mAP: 85.2 -> 85.6 | Helped |
| Use adaptive kernel selection instead of fixed fusion (ablation) | BiCnet · MARS | mAP: 85.5 -> 86.0 | Helped |
| Add a third, 64×32-resolution branch (ablation) | Two-branch network · MARS | mAP: 84.8 -> 79.1 | Hurt |
| Increase the small-to-big frame ratio from α=1 to α=4 (ablation) | Two-branch network · MARS | mAP: 84.8 -> 83.8 | Hurt |
| Add Cross-Scale Paths (ablation) | Two-branch network · MARS | mAP: 0.7% | Helped |
| Add an adaptive TKS block after stage 2 (ablation) | BiCnet · MARS | mAP: 0.4% | Helped |
| Use two temporal kernels of different sizes instead of one (ablation) | TKS · MARS | mAP | Helped |
| Increase the temporal kernel count from two to three (ablation) | TKS · MARS | mAP: 86.0 -> 85.8 | No clear effect |
| Place TKS after stage 1 rather than stage 2 (ablation) | BiCnet · MARS | mAP: 86.0 -> 85.3 | Hurt |
| Place TKS after stage 4 rather than stage 2 (ablation) | BiCnet · MARS | mAP: 86.0 -> 85.4 | Hurt |
Closest papers by meaning
- Spatial-Temporal Graph Convolutional Network for Video-Based Person Re-Identification (CVPR 2020 · similarity 0.90)
- Temporal-Enhanced Convolutional Network for Person Re-Identification (AAAI 2018 · similarity 0.89)
- Diversity Regularized Spatiotemporal Attention for Video-Based Person Re-Identification (CVPR 2018 · similarity 0.89)
- Spatial-Temporal Correlation and Topology Learning for Person Re-Identification in Videos (CVPR 2021 · similarity 0.89)
- Spatial and Temporal Mutual Promotion for Video-Based Person Re-Identification (AAAI 2019 · similarity 0.89)
- TF-CLIP: Learning Text-Free CLIP for Video-Based Person Re-identification (AAAI 2024 · 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.