… We develop a proposal-based 3D video completion method. Different from prior work, we
argue that proposals accurately summarize spatially and temporally non-local candidates that …

Existing approaches for spatio-temporal action detection in videos are limited by the spatial
extent and temporal duration of the actions. In this paper, we present a modular system for …

… proposals, which weigh the original video features to generate proposal-… video and query
features. Subsequently, a multi-scale proposal generation module is designed to refine video …

… : A simple pipeline1 for the conventional proposal-based WSVG methods is illustrated in Fig-
… After the multi-step refinements, our IRON gradually converges to more complete intervals …

… 2019), we design a novel semantic completion module that predicts the important words (eg …
given by semantic completion module, we compute reward for each proposal based on the …

… Video moment retrieval aims to precisely identify and localize a specific segment within an
untrimmed video … , objects, and motions inherent in videos. Additionally, existing approaches …

… : Proposal-based method, Proposal-free method, and Reinforcement Learning-based method.
In early works, some proposalbased … between video and query, treat the video as a whole …

… In this paper, we advocate a novel proposal-based paradigm that … As a result, our
proposal-based sound source localization … Our location map tends to capture larger and more …

… offer a more complete and high-level … proposal-based video saliency detection approach.
The two frames in the first row are the original frames from the horses01 and people05 videos …

… actions in untrimmed videos with only video-level category … are supervised by the labels of
videos. However, the objective for … Step-by-step [64] allows the model to learn more complete …