Video summarization with a convolutional attentive adversarial network

作者：

Highlights：

• We integrate the self-attention mechanism and a fully convolutional sequence network to capture the global and local temporal dependencies of video frames.

• A convolutional attentive generative adversarial network is designed for unsupervised video summarization.

• By discarding recurrent structures, our generator can be paralleled much easier.

• Experiment results show that our method not only achieves better or comparable performance within unsupervised methods, but also is superior to most of the published supervised approaches.

摘要

•We integrate the self-attention mechanism and a fully convolutional sequence network to capture the global and local temporal dependencies of video frames.•A convolutional attentive generative adversarial network is designed for unsupervised video summarization.•By discarding recurrent structures, our generator can be paralleled much easier.•Experiment results show that our method not only achieves better or comparable performance within unsupervised methods, but also is superior to most of the published supervised approaches.

论文关键词：Video summarization,Generative adversarial network,Self attention

论文评审过程：Received 25 November 2020, Revised 5 June 2022, Accepted 9 June 2022, Available online 11 June 2022, Version of Record 22 June 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108840