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

Existing action detection algorithms usually generate action proposalsthrough an extensive search over the video at multiple temporal scales, whichbrings about huge computational overhead and deviates from the human perceptionprocedure. We argue that the process of detecting actions should be naturallyone of observation and refinement: observe the current window and refine thespan of attended window to cover true action regions. In this paper, we proposean active action proposal model that learns to find actions throughcontinuously adjusting the temporal bounds in a self-adaptive way. The wholeprocess can be deemed as an agent, which is firstly placed at a position in thevideo at random, adopts a sequence of transformations on the current attendedregion to discover actions according to a learned policy. We utilizereinforcement learning, especially the Deep Q-learning algorithm to learn theagent's decision policy. In addition, we use temporal pooling operation toextract more effective feature representation for the long temporal window, anddesign a regression network to adjust the position offsets between predictedresults and the ground truth. Experiment results on THUMOS 2014 validate theeffectiveness of the proposed approach, which can achieve competitiveperformance with current action detection algorithms via much fewer proposals.

A Self-Adaptive Proposal Model for Temporal Action Detection based on Reinforcement Learning