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

Unsupervised domain adaptation which aims to adapt models trained on alabeled source domain to a completely unlabeled target domain has attractedmuch attention in recent years. While many domain adaptation techniques havebeen proposed for images, the problem of unsupervised domain adaptation invideos remains largely underexplored. In this paper, we introduce Contrast andMix (CoMix), a new contrastive learning framework that aims to learndiscriminative invariant feature representations for unsupervised video domainadaptation. First, unlike existing methods that rely on adversarial learningfor feature alignment, we utilize temporal contrastive learning to bridge thedomain gap by maximizing the similarity between encoded representations of anunlabeled video at two different speeds as well as minimizing the similaritybetween different videos played at different speeds. Second, we propose a novelextension to the temporal contrastive loss by using background mixing thatallows additional positives per anchor, thus adapting contrastive learning toleverage action semantics shared across both domains. Moreover, we alsointegrate a supervised contrastive learning objective using targetpseudo-labels to enhance discriminability of the latent space for video domainadaptation. Extensive experiments on several benchmark datasets demonstrate thesuperiority of our proposed approach over state-of-the-art methods. Projectpage: https://cvir.github.io/projects/comix

Contrast and Mix: Temporal Contrastive Video Domain Adaptation with Background Mixing