SUSiNet: See, Understand and Summarize it

Abstract

In this work we propose a multi-task spatio-temporal network, called SUSiNet, that can jointly tackle the spatio-temporal problems of saliency estimation, action recognition and video summarization. Our approach employs a single network that is jointly end-to-end trained for all tasks with multiple and diverse datasets related to the exploring tasks. The proposed network uses a unified architecture that includes global and task specific layer and produces multiple output types, i.e., saliency maps or classification labels, by employing the same video input. Moreover, one additional contribution is that the proposed network can be deeply supervised through an attention module that is related to human attention as it is expressed by eye-tracking data. From the extensive evaluation, on seven different datasets, we have observed that the multi-task network performs as well as the state-of-the-art single-task methods (or in some cases better), while it requires less computational budget than having one independent network per each task.

arXiv commentCVPR Workshops 2019 (Mutual benefits of cognitive and computer vision)

References in corpus (10)

• Two-Stream Convolutional Networks for Action Recognition in Videos
• Action Recognition with Trajectory-Pooled Deep-Convolutional Descriptors
• Spatiotemporal Residual Networks for Video Action Recognition
• Predicting Video Saliency with Object-to-Motion CNN and Two-layer Convolutional LSTM
• Convolutional Oriented Boundaries
• Shallow and Deep Convolutional Networks for Saliency Prediction
• Recurrent Mixture Density Network for Spatiotemporal Visual Attention
• Revisiting Video Saliency: A Large-scale Benchmark and a New Model
• Multimodal Visual Concept Learning with Weakly Supervised Techniques
• End-to-End Saliency Mapping via Probability Distribution Prediction