Research map / Video / Video segmentation and tracking
Video-Based Person Re-Identification With Spatial and Temporal Memory Networks
Chanho Eom, Geon Lee, Junghyup Lee, Bumsub Ham
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 spatial memory to refine frame features (ablation) | video re-ID baseline · MARS | rank-1 / mAP: 87.3 -> 89.3 rank-1, 79.1 -> 82.5 mAP | Improved |
| add spatial memory to refine frame features (ablation) | video re-ID baseline · LS-VID | rank-1 / mAP: 71.6 -> 79.6 rank-1, 55.9 -> 65.8 mAP | Improved |
| add spatial memory to refine frame features (ablation) | video re-ID baseline · DukeMTMC-VideoReID | rank-1 / mAP: 95.0 -> 96.2 rank-1, 92.7 -> 94.2 mAP | Improved |
| add temporal memory to aggregate frame features (ablation) | video re-ID baseline · MARS | rank-1 / mAP: 87.3 -> 89.5 rank-1, 79.1 -> 82.0 mAP | Improved |
| add temporal memory to aggregate frame features (ablation) | video re-ID baseline · LS-VID | rank-1 / mAP: 71.6 -> 78.9 rank-1, 55.9 -> 64.4 mAP | Improved |
| add temporal memory to aggregate frame features (ablation) | video re-ID baseline · DukeMTMC-VideoReID | rank-1 / mAP: 95.0 -> 95.4 rank-1, 92.7 -> 93.7 mAP | Improved |
| add temporal memory on top of spatial memory (ablation) | spatial memory model · MARS | rank-1 / mAP: 89.3 -> 89.9 rank-1, 82.5 -> 83.7 mAP | Improved |
| add temporal memory on top of spatial memory (ablation) | spatial memory model · LS-VID | rank-1 / mAP: 79.6 -> 80.6 rank-1, 65.8 -> 66.6 mAP | Improved |
| add memory spread loss to both memories (ablation) | STMN · MARS | rank-1 / mAP: 89.1 -> 89.9 rank-1, 81.9 -> 83.7 mAP | Improved |
| add memory spread loss to both memories (ablation) | STMN · LS-VID | rank-1 / mAP: 79.9 -> 80.6 rank-1, 65.4 -> 66.6 mAP | Improved |
| add memory spread loss to the spatial memory only (ablation) | baseline + spatial memory · MARS | rank-1 / mAP: 88.7 -> 89.3 rank-1, 81.6 -> 82.5 mAP | Improved |
| add memory spread loss to the temporal memory only (ablation) | baseline + temporal memory · MARS | rank-1 / mAP: 88.5 -> 89.5 rank-1, 81.9 -> 82.0 mAP | Mixed |
Closest papers by meaning
- Spatial and Temporal Mutual Promotion for Video-Based Person Re-Identification (AAAI 2019 · similarity 0.91)
- Spatio-Temporal Representation Factorization for Video-Based Person Re-Identification (ICCV 2021 · similarity 0.91)
- BiCnet-TKS: Learning Efficient Spatial-Temporal Representation for Video Person Re-Identification (CVPR 2021 · similarity 0.91)
- Relation-Guided Spatial Attention and Temporal Refinement for Video-Based Person Re-Identification (AAAI 2020 · similarity 0.91)
- TF-CLIP: Learning Text-Free CLIP for Video-Based Person Re-identification (AAAI 2024 · similarity 0.89)
- CLIMB-ReID: A Hybrid CLIP-Mamba Framework for Person Re-Identification (AAAI 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.
