Research map / Continual learning / Domain and prompt-based
Bi-C 2 R: Bidirectional Continual Compatible Representation for Re-Indexing Free Lifelong Person Re-Identification
Zhenyu Cui, Jiahuan Zhou, Yuxin Peng
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 Bi-C2R on top of LSTKC for re-indexing-free lifelong ReID (ablation) | LSTKC · Order-1 RFL-ReID | average mAP / R1: 47.1 -> 50.1 mAP, 58.6 -> 61.5 R1 | Improved |
| add BiCT and BiCD to LSTKC base (ablation) | LSTKC · Order-1 RFL-ReID | mAP: 47.1 -> 48.3 | Improved |
| add BiAD to LSTKC + BiCT + BiCD (ablation) | Bi-C2R · Order-1 RFL-ReID | mAP: 48.3 -> 49.0 | Improved |
| add dynamic feature fusion (DFF) (ablation) | Bi-C2R · Order-1 RFL-ReID | mAP: 49.0 -> 50.1 | Improved |
| add BiAD (ablation) | Bi-C2R · Order-1 L-ReID | mAP: 51.0 -> 51.4 | Improved |
| add DFF (ablation) | Bi-C2R · Order-1 L-ReID | mAP: 51.4 -> 51.4 | No clear effect |
| use adaptive balancing factor instead of fixed a^m values 0.0 to 1.0 in BiCT (ablation) | BiCT network · Order-1 RFL-ReID | mAP: 49.8 -> 50.1 | Improved |
| use both FMM and KCM instead of FMM only (ablation) | BiCT network · Order-1 RFL-ReID, average | mAP: 49.6 -> 50.1 | Improved |
| use BiCT-Net instead of Full-Conv transfer network (ablation) | BiCT network · Order-1 RFL-ReID | mAP: 48.9 -> 50.1 | Improved |
| use DFF instead of time-increasing fusion weight (ablation) | Bi-C2R · Order-1 RFL-ReID | mAP: 49.2 -> 50.1 | Improved |
| use Bi-C2R instead of C2R | C2R · Order-1 average forgetting | AF mAP/R1: 10.6/11.1 -> 7.6/7.1 | Improved |
| use feature updating instead of re-indexing gallery | Bi-C2R · inference time per stage | speedup: over x20 acceleration | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- LSTKC — LSTKC: Long Short-Term Knowledge Consolidation for Lifelong Person Re-identification (AAAI 2024) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Learning Continual Compatible Representation for Re-indexing Free Lifelong Person Re-identification (CVPR 2024 · similarity 0.96)
- A Memorizing and Generalizing Framework for Lifelong Person Re-Identification (TPAMI 2023 · similarity 0.90)
- Instruct-ReID: A Multi-purpose Person Re-identification Task with Instructions (CVPR 2024 · similarity 0.87)
- DKC: Differentiated Knowledge Consolidation for Cloth-Hybrid Lifelong Person Re-identification (CVPR 2025 · similarity 0.87)
- LSTKC: Long Short-Term Knowledge Consolidation for Lifelong Person Re-identification (AAAI 2024 · similarity 0.87)
- Cheb-GR: Rethinking K-nearest Neighbor Search in Re-ranking for Person Re-identification (CVPR 2025 · similarity 0.87)
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.
