FAR: Fourier Aerial Video Recognition

TL;DR

FAR introduces a Fourier-based method for UAV video activity recognition that effectively disentangles human actions from backgrounds and models spatial-temporal dependencies with reduced computation, achieving significant accuracy improvements.

Contribution

The paper proposes a novel Fourier object disentanglement and Fourier Attention mechanism for UAV video recognition, enhancing accuracy and efficiency over existing methods.

Findings

Achieves 8.02% to 38.69% higher top-1 accuracy on UAV datasets.

Operates up to 3 times faster than prior methods.

Effectively separates human actions from backgrounds in aerial videos.

Abstract

We present an algorithm, Fourier Activity Recognition (FAR), for UAV video activity recognition. Our formulation uses a novel Fourier object disentanglement method to innately separate out the human agent (which is typically small) from the background. Our disentanglement technique operates in the frequency domain to characterize the extent of temporal change of spatial pixels, and exploits convolution-multiplication properties of Fourier transform to map this representation to the corresponding object-background entangled features obtained from the network. To encapsulate contextual information and long-range space-time dependencies, we present a novel Fourier Attention algorithm, which emulates the benefits of self-attention by modeling the weighted outer product in the frequency domain. Our Fourier attention formulation uses much fewer computations than self-attention. We have evaluated our approach on multiple UAV datasets including UAV Human RGB, UAV Human Night, Drone Action, and NEC Drone. We demonstrate a relative improvement of 8.02% - 38.69% in top-1 accuracy and up to 3 times faster over prior works.