AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Patch-Based Discriminative Feature Learning for Unsupervised Person Re-Identification

Qize Yang, Hong-Xing Yu, Ancong Wu, Wei-Shi Zheng

CVPR 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Supervised person re-identification · 255 citations (OpenAlex)

Check 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 changedWhereMeasuredEffect
Learn patch features with PEDAL rather than using the PatchNet baseline alone (ablation)PAUL · Market-1501Rank-1: 59.3 -> 66.2Improved
Learn patch features with PEDAL rather than using the PatchNet baseline alone (ablation)PAUL · DukeMTMC-reIDmAP: 45.6 -> 52.1Improved
Add image-level IPFL to the PatchNet baseline (ablation)PAUL · Market-1501Rank-1: 59.3 -> 65.4Improved
Add image-level IPFL to the PatchNet baseline (ablation)PAUL · DukeMTMC-reIDRank-1: 65.7 -> 67.1Improved
Use both PEDAL and IPFL rather than either loss alone (ablation)PAUL · Market-1501mAP: 38.0 -> 40.1Improved
Use both PEDAL and IPFL rather than either loss alone (ablation)PAUL · DukeMTMC-reIDmAP: 52.1 -> 53.2Improved
Learn patch-level features rather than use image-level ResNet-50 features (ablation)PAUL · Market-1501Rank-1: 46.6 -> 59.3Improved
Apply PEDAL to patches rather than pull similar whole-image features together (ablation)PAUL · Market-1501Rank-1: 25.6 -> 66.2Improved
Use learnable PGN patch selection rather than random patch selection (ablation)PAUL · Market-1501Rank-1: 24.1 -> 68.5Improved
Use learnable PGN patch selection rather than equal horizontal stripes (ablation)PAUL · DukeMTMC-reIDRank-1: 56.2 -> 72.0Improved
Use 6 patches instead of 4 (ablation)PAUL · Market-1501Rank-1: 67.3 -> 68.5Improved
Use 8 patches instead of 6 (ablation)PAUL · Market-1501mAP: 40.1 -> 37.2Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy