Research

AR-Net: Adaptive Frame Resolution for Efficient Action Recognition

ECCV

Published on

07/31/2020

Categories

Computer Vision ECCV

Action recognition is an open and challenging problem in computer vision. While current state-of-the-art models offer excellent recognition results, their computational expense limits their impact for many real-world applications. In this paper, we propose a novel approach, called AR-Net (Adaptive Resolution Network), that selects on-the-fly the optimal resolution for each frame conditioned on the input for efficient action recognition in long untrimmed videos. Specifically, given a video frame, a policy network is used to decide what input resolution should be used for processing by the action recognition model, with the goal of improving both accuracy and efficiency. We efficiently train the policy network jointly with the recognition model using standard backpropagation. Extensive experiments on several challenging action recognition benchmark datasets well demonstrate the efficacy of our proposed approach over state-of-the-art methods. The project page can be found at https://mengyuest.github.io/AR-Net.

This paper has been published at ECCV 2020

Please cite our work using the BibTeX below.

@misc{meng2020arnet,
      title={AR-Net: Adaptive Frame Resolution for Efficient Action Recognition}, 
      author={Yue Meng and Chung-Ching Lin and Rameswar Panda and Prasanna Sattigeri and Leonid Karlinsky and Aude Oliva and Kate Saenko and Rogerio Feris},
      year={2020},
      eprint={2007.15796},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
Close Modal