Research map / Video / Video segmentation and tracking
Learning Modal-Invariant and Temporal-Memory for Video-Based Visible-Infrared 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 modal-invariant adversarial learning (ablation) | Baseline · HITSZ-VCM infrared to visible | Rank-1: 55.58 -> 59.73 | Helped |
| add modal-invariant adversarial learning (ablation) | Baseline · HITSZ-VCM visible to infrared | mAP: 42.61 -> 43.59 | Helped |
| add Temporal Memory Refinement (ablation) | Baseline · HITSZ-VCM infrared to visible | mAP: 40.80 -> 43.87 | Helped |
| add Temporal Memory Refinement (ablation) | Baseline · HITSZ-VCM visible to infrared | Rank-1: 59.58 -> 62.19 | Helped |
| shuffle frames in TMR (ablation) | MITML · HITSZ-VCM infrared to visible | Rank-1: 63.74 -> 60.82 | Hurt |
| shuffle frames in TMR (ablation) | MITML · HITSZ-VCM visible to infrared | mAP: 47.69 -> 45.64 | Hurt |
| replace MITML modal-invariant learning with cmGAN adversarial learning (ablation) | MITML · HITSZ-VCM infrared to visible | Rank-1: 57.96 -> 63.74 | Hurt |
| replace MITML modal-invariant learning with UCDA adversarial learning (ablation) | MITML · HITSZ-VCM visible to infrared | Rank-1: 64.15 -> 64.54 | Hurt |
| vary the weight λ of the modal-invariant loss from 0.01 to 1.0 (ablation) | MITML · HITSZ-VCM infrared to visible | mAP: 45.54; 46.50; 46.19; 46.49; 47.32; 45.71; 45.46; 45.31 | Mixed |
| use max pooling instead of average pooling (ablation) | Baseline · HITSZ-VCM infrared to visible | Rank-1: 55.58 -> 54.47 | Hurt |
| use weighted average pooling instead of average pooling (ablation) | Baseline · HITSZ-VCM infrared to visible | mAP: 40.80 -> 36.52 | Hurt |
| increase the number of frames in a tracklet from one to six (ablation) | Baseline · HITSZ-VCM | mAP: 23.09% | Helped |
Closest papers by meaning
- Cross-Modality Paired-Images Generation for RGB-Infrared Person Re-Identification (AAAI 2020 · similarity 0.91)
- Robust Multi-Modality Person Re-identification (AAAI 2021 · similarity 0.90)
- Infrared-Visible Cross-Modal Person Re-Identification with an X Modality (AAAI 2020 · similarity 0.89)
- Neural Feature Search for RGB-Infrared Person Re-Identification (CVPR 2021 · similarity 0.89)
- TokenMatcher: Diverse Tokens Matching for Unsupervised Visible-Infrared Person Re-Identification (AAAI 2025 · similarity 0.88)
- Diffusion-based Synthetic Data Generation for Visible-Infrared 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.