Research map / Video / Video segmentation and tracking
Multi-Granularity Reference-Aided Attentive Feature Aggregation for Video-Based 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 |
|---|---|---|---|
| replace global spatial-temporal average pooling with MG-RAFA | ResNet-50 video re-identification · MARS | mAP: 82.1 -> 85.9 | Helped |
| replace global spatial-temporal average pooling with MG-RAFA | ResNet-50 video re-identification · iLIDS-VID | Rank-1: 86.5 -> 88.6 | Helped |
| replace global spatial-temporal average pooling with MG-RAFA | ResNet-50 video re-identification · PRID2011 | Rank-1: 92.5 -> 95.9 | Helped |
| add global reference-node relations to appearance-only attention (ablation) | MG-AFA (N=4) · MARS | Rank-1: 86.6 -> 88.8 | Helped |
| add global reference-node relations to appearance-only attention (ablation) | MG-AFA (N=4) · iLIDS-VID | Rank-1: 86.7 -> 88.6 | Helped |
| add global reference-node relations to appearance-only attention (ablation) | MG-AFA (N=4) · PRID2011 | Rank-1: 92.6 -> 95.9 | Helped |
| add multiple granularity levels instead of using a single granularity (ablation) | SG-RAFA (S=1) · MARS | Rank-1: 87.8 -> 88.8 | Helped |
| add multiple granularity levels instead of using a single granularity (ablation) | SG-RAFA (S=1) · iLIDS-VID | Rank-1: 87.1 -> 88.6 | Helped |
| add multiple granularity levels instead of using a single granularity (ablation) | SG-RAFA (S=1) · PRID2011 | Rank-1: 93.6 -> 95.9 | Helped |
| increase the granularity count from two to four (ablation) | MG-RAFA · MARS | mAP: 85.5 -> 85.9 | Helped |
| use channel-split attention at one shared granularity rather than different granularities (ablation) | SG-RAFA (S=4) · MARS | mAP: 84.9 -> 85.9 | Hurt |
| average-pool reference nodes across time to form a representative set (ablation) | SG-RAFA · MARS | mAP: 84.3 -> 85.1 | Helped |
Closest papers by meaning
- Learning Multi-Granular Hypergraphs for Video-Based Person Re-Identification (CVPR 2020 · similarity 0.88)
- Diversity Regularized Spatiotemporal Attention for Video-Based Person Re-Identification (CVPR 2018 · similarity 0.88)
- Relation-Guided Spatial Attention and Temporal Refinement for Video-Based Person Re-Identification (AAAI 2020 · similarity 0.88)
- X-ReID: Multi-granularity Information Interaction for Video-Based Visible-Infrared Person Re-Identification (AAAI 2026 · similarity 0.87)
- Attribute-Driven Feature Disentangling and Temporal Aggregation for Video Person Re-Identification (CVPR 2019 · similarity 0.87)
- 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.