Capturing Temporal Information in a Single Frame: Channel Sampling Strategies for Action Recognition
This work addresses the challenge of efficient temporal modeling in video understanding for the computer vision community, offering a flexible and low-cost solution that is incremental but widely applicable.
The paper tackles the problem of capturing temporal information for video classification in 2D networks without increasing computational cost by proposing a novel channel sampling strategy that reorders input video channels to capture short-term frame-to-frame changes, resulting in performance improvements of up to 24% over baselines on multiple architectures and datasets.
We address the problem of capturing temporal information for video classification in 2D networks, without increasing their computational cost. Existing approaches focus on modifying the architecture of 2D networks (e.g. by including filters in the temporal dimension to turn them into 3D networks, or using optical flow, etc.), which increases computation cost. Instead, we propose a novel sampling strategy, where we re-order the channels of the input video, to capture short-term frame-to-frame changes. We observe that without bells and whistles, the proposed sampling strategy improves performance on multiple architectures (e.g. TSN, TRN, TSM, and MVFNet) and datasets (CATER, Something-Something-V1 and V2), up to 24% over the baseline of using the standard video input. In addition, our sampling strategies do not require training from scratch and do not increase the computational cost of training and testing. Given the generality of the results and the flexibility of the approach, we hope this can be widely useful to the video understanding community. Code is available on our website: https://github.com/kiyoon/channel_sampling.