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 many real-world problems, there is typically a large discrepancy betweenthe characteristics of data used in training versus deployment. A prime exampleis the analysis of aggression videos: in a criminal incidence, typicallysuspects need to be identified based on their clean portrait-like photos,instead of their prior video recordings. This results in three majorchallenges; large domain discrepancy between violence videos and ID-photos, thelack of video examples for most individuals and limited training dataavailability. To mimic such scenarios, we formulate a realistic domain-transferproblem, where the goal is to transfer the recognition model trained on cleanposed images to the target domain of violent videos, where training videos areavailable only for a subset of subjects. To this end, we introduce theWildestFaces dataset, tailored to study cross-domain recognition under avariety of adverse conditions. We divide the task of transferring a recognitionmodel from the domain of clean images to the violent videos into twosub-problems and tackle them using (i) stacked affine-transforms forclassifier-transfer, (ii) attention-driven pooling for temporal-adaptation. Weadditionally formulate a self-attention based model for domain-transfer. Weestablish a rigorous evaluation protocol for this clean-to-violent recognitiontask, and present a detailed analysis of the proposed dataset and the methods.Our experiments highlight the unique challenges introduced by the WildestFacesdataset and the advantages of the proposed approach.

Red Carpet to Fight Club: Partially-supervised Domain Transfer for Face Recognition in Violent Videos