AI Paper Atlas

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

Interpretable and Generalizable Person Re-Identification with Query-Adaptive Convolution and Temporal Lifting

Shengcai Liao, Ling Shao

ECCV 2020 · Label-efficient and robust learning › Recognition and long-tail · approach: Supervised person re-identification · 180 citations (OpenAlex)

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 changedWhereMeasuredEffect
use QAConv matching instead of softmax cross-entropy (ablation)ResNet-152 re-ID · Market-1501 -> DukeMTMC-reIDRank-1 (%): 34.9 -> 54.4Improved
use QAConv matching instead of softmax cross-entropy (ablation)ResNet-152 re-ID · DukeMTMC-reID -> Market-1501Rank-1 (%): 48.5 -> 62.8Improved
use QAConv matching instead of class memory loss with plain features (ablation)ResNet-152 re-ID · Market-1501 -> DukeMTMC-reIDmAP (%): 21.8 -> 33.6Improved
add Arc loss margin instead of softmax (ablation)ResNet-152 re-ID · Market-1501 -> DukeMTMC-reIDmAP (%): 18.4 -> 17.1No clear effect
add re-ranking (ablation)QAConv · Market-1501 -> DukeMTMC-reIDmAP (%): 33.6 -> 52.4Improved
add re-ranking (ablation)QAConv · DukeMTMC-reID -> Market-1501mAP (%): 31.6 -> 51.2Improved
add TLift on top of re-ranking (ablation)QAConv + RR · Market-1501 -> DukeMTMC-reIDmAP (%): 52.4 -> 61.2Improved
add TLift on top of re-ranking (ablation)QAConv + RR · DukeMTMC-reID -> Market-1501Rank-1 (%): 68.5 -> 78.7Improved
add TLift without re-ranking (ablation)QAConv · DukeMTMC-reID -> Market-1501Rank-1 (%): 62.8 -> 61.5Mixed
use TFusion instead of TLift after re-ranking (ablation)QAConv + RR · DukeMTMC-reID -> Market-1501mAP (%): 51.2 -> 47.2Hurt
use ResNet-152 backbone instead of ResNet-50 (ablation)QAConv · Market-1501 -> DukeMTMC-reIDRank-1 (%): 48.8 -> 54.4Improved
use ResNet-152 backbone instead of ResNet-50 (ablation)QAConv · DukeMTMC-reID -> Market-1501mAP (%): 27.2 -> 31.6Improved

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