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

In visible-infrared video person re-identification (re-ID), extractingfeatures not affected by complex scenes (such as modality, camera views,pedestrian pose, background, etc.) changes, and mining and utilizing motioninformation are the keys to solving cross-modal pedestrian identity matching.To this end, the paper proposes a new visible-infrared video person re-IDmethod from a novel perspective, i.e., adversarial self-attack defense andspatial-temporal relation mining. In this work, the changes of views, posture,background and modal discrepancy are considered as the main factors that causethe perturbations of person identity features. Such interference informationcontained in the training samples is used as an adversarial perturbation. Itperforms adversarial attacks on the re-ID model during the training to make themodel more robust to these unfavorable factors. The attack from the adversarialperturbation is introduced by activating the interference information containedin the input samples without generating adversarial samples, and it can be thuscalled adversarial self-attack. This design allows adversarial attack anddefense to be integrated into one framework. This paper further proposes aspatial-temporal information-guided feature representation network to use theinformation in video sequences. The network cannot only extract the informationcontained in the video-frame sequences but also use the relation of the localinformation in space to guide the network to extract more robust features. Theproposed method exhibits compelling performance on large-scale cross-modalityvideo datasets. The source code of the proposed method will be released athttps://github.com/lhf12278/xxx.

Adversarial Self-Attack Defense and Spatial-Temporal Relation Mining for Visible-Infrared Video Person Re-Identification