Research map / Label-efficient and robust learning / Self-supervised and contrastive
Self-Supervised Geometric Features Discovery via Interpretable Attention for Vehicle Re-Identification and Beyond
Ming Li, Xinming Huang, Ziming Zhang
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 the interpretable-attention geometric branch (GFB, K=7) (ablation) | GB baseline · VeRi-776 | tmAP (%): 84.0 -> 85.9 | Improved |
| add the interpretable-attention geometric branch (GFB, K=7) (ablation) | GB baseline · CityFlow-ReID | imAP (%): 32.04 -> 36.63 | Improved |
| add a plain ResNet18 branch without attention (ablation) | GB baseline · CityFlow-ReID | imAP (%): 32.04 -> 34.63 | Improved |
| use IAM-based GFB instead of a plain ResNet18 branch (ablation) | GB baseline · CityFlow-ReID | imAP (%): 34.63 -> 36.63 | Improved |
| add self-supervised rotation learning (SLB, lambda 1.0) to GB+GFB (ablation) | GB+GFB · VeRi-776 | tmAP (%): 85.9 -> 86.2 | Improved |
| add self-supervised rotation learning (SLB, lambda 1.0) to GB+GFB (ablation) | GB+GFB · CityFlow-ReID | imAP (%): 36.63 -> 37.14 | Improved |
| add self-supervised rotation learning (SLB, lambda 1.0) to GB+GFB (ablation) | GB+GFB · CityFlow-ReID | Top-1 (%): 59.98 -> 60.08 | Improved |
| increase attention neighborhood K from 7 to 15 (ablation) | GB+GFB · CityFlow-ReID | imAP (%): 36.63 -> 36.32 | No clear effect |
| raise lambda_SLB from 0.1 to 1.0 (ablation) | GB+GFB+SLB · VeRi-776 | tmAP (%): 85.8 -> 86.2 | No clear effect |
| raise lambda_SLB from 1.0 to 2.0 (ablation) | GB+GFB+SLB · CityFlow-ReID | imAP (%): 37.14 -> 36.54 | No clear effect |
| use ours instead of SAVER (no extra supervision) | Ours · VehicleID Small gallery | Top-1 (%): 79.9 -> 86.8 | Improved |
| use ours instead of PVEN (uses extra supervision) | Ours · VeRi-776 | imAP (%): 79.5 -> 81.0 | Improved |
Closest papers by meaning
- The Devil is in the Details: Self-Supervised Attention for Vehicle Re-Identification (ECCV 2020 · similarity 0.90)
- Disentangled Feature Learning Network and a Comprehensive Benchmark for Vehicle Re-Identification (TPAMI 2021 · similarity 0.89)
- Orientation Invariant Feature Embedding and Spatial Temporal Regularization for Vehicle Re-Identification (ICCV 2017 · similarity 0.88)
- Learning Coarse-to-Fine Structured Feature Embedding for Vehicle Re-Identification (AAAI 2018 · similarity 0.87)
- Infer the Whole from a Glimpse of a Part: Keypoint-Based Knowledge Graph for Vehicle Re-Identification (AAAI 2025 · similarity 0.86)
- CLIP-driven View-aware Prompt Learning for Unsupervised Vehicle Re-identification (AAAI 2025 · similarity 0.84)
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.
