Shi Pu

TL;DR: In this article, a reciprocative learning algorithm was proposed to exploit visual attention for training deep classifiers, which consists of feed-forward and backward operations to generate attention maps, which serve as regularization terms coupled with the original classification loss function for training.

TL;DR: In this paper, a reciprocative learning algorithm was proposed to exploit visual attention for training deep classifiers, which consists of feed-forward and backward operations to generate attention maps, which serve as regularization terms coupled with the original classification loss function for training.

TL;DR: An end-to-end model via Information Tree for One-Shot video grounding (IT-OS), which can eliminate the interference of irrelevant frames based on branch search and branch cropping techniques and several self-supervised tasks are proposed based on the information tree to improve the representation learning under insufficient labeling.

TL;DR: This paper introduces the weakly-supervised contrastive learning that classifies the video as action-consistent and action-independent frames relying on the video-caption action semantic correspondence and designs the hierarchical reasoning layers to decouple fully connected multi-head attention and remove the redundant interfering correlations.

TL;DR: A recurrent memory activation network (RMAN) to exploit the untapped temporal coherence of the target appearance for visual tracking, built on top of the long short-term memory network (LSTM) with an additional memory activation layer.