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

The goal of weakly supervised video anomaly detection is to learn a detectionmodel using only video-level labeled data. However, prior studies typicallydivide videos into fixed-length segments without considering the complexity orduration of anomalies. Moreover, these studies usually just detect the mostabnormal segments, potentially overlooking the completeness of anomalies. Toaddress these limitations, we propose a Dynamic Erasing Network (DE-Net) forweakly supervised video anomaly detection, which learns multi-scale temporalfeatures. Specifically, to handle duration variations of abnormal events, wefirst propose a multi-scale temporal modeling module, capable of extractingfeatures from segments of varying lengths and capturing both local and globalvisual information across different temporal scales. Then, we design a dynamicerasing strategy, which dynamically assesses the completeness of the detectedanomalies and erases prominent abnormal segments in order to encourage themodel to discover gentle abnormal segments in a video. The proposed methodobtains favorable performance compared to several state-of-the-art approacheson three datasets: XD-Violence, TAD, and UCF-Crime. Code will be made availableat https://github.com/ArielZc/DE-Net.

Dynamic Erasing Network Based on Multi-Scale Temporal Features for Weakly Supervised Video Anomaly Detection