Agent-Aware Boundary Network for Temporal Action Proposal Generation
ABN: Agent-Aware Boundary Networks for Temporal Action Proposal Generation
In this paper, we propose a novel framework named agent-awareboundary network (abn), which consists of two sub-networks (i)an agent-aware representation network to obtain both agent-agent and agents-environmentrelationships in the video representation, and (ii)a boundary generationnetwork to estimate the confidence score of temporal intervals.Comprehensive evaluations on 20-actionthumos-14 and 200-action activitynet-1.3 datasets with different backbone networks show that our proposed abn robustly outperforms state-of-the-art methods regardless of the employed backbonenetwork on temporal action proposal generation (tapg).We further examine the proposal quality by leveraging the proposal generated by our method onto temporal action detection (tad)frameworks and evaluate their detection performances.
