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

Learning modality invariant features is central to the problem ofVisible-Thermal cross-modal Person Reidentification (VT-ReID), where query andgallery images come from different modalities. Existing works implicitly alignthe modalities in pixel and feature spaces by either using adversarial learningor carefully designing feature extraction modules that heavily rely on domainknowledge. We propose a simple but effective framework, MMD-ReID, that reducesthe modality gap by an explicit discrepancy reduction constraint. MMD-ReIDtakes inspiration from Maximum Mean Discrepancy (MMD), a widely usedstatistical tool for hypothesis testing that determines the distance betweentwo distributions. MMD-ReID uses a novel margin-based formulation to matchclass-conditional feature distributions of visible and thermal samples tominimize intra-class distances while maintaining feature discriminability.MMD-ReID is a simple framework in terms of architecture and loss formulation.We conduct extensive experiments to demonstrate both qualitatively andquantitatively the effectiveness of MMD-ReID in aligning the marginal and classconditional distributions, thus learning both modality-independent andidentity-consistent features. The proposed framework significantly outperformsthe state-of-the-art methods on SYSU-MM01 and RegDB datasets. Code will bereleased at https://github.com/vcl-iisc/MMD-ReID

MMD-ReID: A Simple but Effective Solution for Visible-Thermal Person ReID