English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Occluded person re-identification (Re-ID) aims to address the potentialocclusion problem when matching occluded or holistic pedestrians from differentcamera views. Many methods use the background as artificial occlusion and relyon attention networks to exclude noisy interference. However, the significantdiscrepancy between simple background occlusion and realistic occlusion cannegatively impact the generalization of the network.To address this issue, wepropose a novel transformer-based Attention Disturbance and Dual-PathConstraint Network (ADP) to enhance the generalization of attention networks.Firstly, to imitate real-world obstacles, we introduce an Attention DisturbanceMask (ADM) module that generates an offensive noise, which can distractattention like a realistic occluder, as a more complex form ofocclusion.Secondly, to fully exploit these complex occluded images, we developa Dual-Path Constraint Module (DPC) that can obtain preferable supervisioninformation from holistic images through dual-path interaction. With ourproposed method, the network can effectively circumvent a wide variety ofocclusions using the basic ViT baseline. Comprehensive experimental evaluationsconducted on person re-ID benchmarks demonstrate the superiority of ADP overstate-of-the-art methods.

Attention Disturbance and Dual-Path Constraint Network for Occluded Person Re-Identification