Moment Boundary Prediction

Given a video, video grounding aims to re- trieve a temporal moment that semantically corresponds to a language query. In this work, we propose a Parallel Attention Network with Sequence …...