Research map / Video / Video segmentation and tracking
Rethinking Temporal Fusion for Video-Based Person Re-Identification on Semantic and Time Aspect
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 |
|---|---|---|---|
| Use multi-stage fusion instead of single-stage late fusion (ablation) | Video-based ReID · MARS | MAP: 75.4 -> 75.6 | Helped |
| Add semantic attention to multi-stage fusion (ablation) | MS Fusion · MARS | MAP: 75.6 -> 77.7 | Helped |
| Use intra-frame attention instead of average pooling (ablation) | Late-fusion temporal aggregation · MARS | MAP: 75.4 -> 76.6 | Helped |
| Use Euclidean inter-frame attention instead of average pooling (ablation) | Late-fusion temporal aggregation · MARS | MAP: 75.4 -> 75.7 | Helped |
| Use Relation Network inter-frame attention instead of average pooling (ablation) | Late-fusion temporal aggregation · MARS | MAP: 75.4 -> 76.8 | Helped |
| Combine intra-frame and Euclidean inter-frame attention (ablation) | Late-fusion temporal aggregation · MARS | MAP: 76.6 -> 79.1 | Helped |
| Use Relation Network rather than Euclidean distance for inter-frame attention (ablation) | Intra/inter-frame attention · MARS | MAP: 79.1 -> 82.4 | Helped |
| Add multi-stage fusion to intra/inter-frame Relation Network attention (ablation) | Intra/Inter-frame (RN) · MARS | MAP: 82.4 -> 85.2 | Helped |
| Use intra/inter-frame Relation Network attention and multi-stage fusion instead of VRSTC | Video-based ReID · MARS | MAP: 82.3 -> 85.2 | Helped |
| Use intra/inter-frame Relation Network attention and multi-stage fusion instead of VRSTC | Video-based ReID · MARS | top1: 88.5 -> 87.1 | Hurt |
| Use the proposed multi-stage temporal fusion method instead of Attribute Driven | Video-based ReID · PRID2011 | top1: 93.9 -> 95.8 | Helped |
| Use the proposed multi-stage temporal fusion method instead of Attribute Driven | Video-based ReID · i-LIDS-VID | top1: 86.3 -> 87.7 | Helped |
Closest papers by meaning
- Relation-Guided Spatial Attention and Temporal Refinement for Video-Based Person Re-Identification (AAAI 2020 · similarity 0.89)
- Attribute-Driven Feature Disentangling and Temporal Aggregation for Video Person Re-Identification (CVPR 2019 · similarity 0.88)
- Spatial and Temporal Mutual Promotion for Video-Based Person Re-Identification (AAAI 2019 · similarity 0.88)
- TF-CLIP: Learning Text-Free CLIP for Video-Based Person Re-identification (AAAI 2024 · similarity 0.87)
- Temporal Correlation Vision Transformer for Video Person Re-Identification (AAAI 2024 · similarity 0.86)
- CLIMB-ReID: A Hybrid CLIP-Mamba Framework for Person Re-Identification (AAAI 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.