The THUMOS Challenge on Action Recognition for Videos "in the Wild"

TL;DR

The THUMOS challenge advances action recognition in videos by introducing untrimmed videos, providing a benchmark for classification and detection, and evaluating methods' generalization from trimmed to untrimmed videos.

Contribution

This paper details the THUMOS benchmark, evaluation protocols, and presents an empirical study on recognizing actions in untrimmed videos, highlighting challenges and future directions.

Findings

Untrimmed videos pose more realistic challenges for action recognition.

Methods trained on trimmed videos have limited generalization to untrimmed videos.

The benchmark facilitates comparison of diverse approaches in action detection.

Abstract

Automatically recognizing and localizing wide ranges of human actions has crucial importance for video understanding. Towards this goal, the THUMOS challenge was introduced in 2013 to serve as a benchmark for action recognition. Until then, video action recognition, including THUMOS challenge, had focused primarily on the classification of pre-segmented (i.e., trimmed) videos, which is an artificial task. In THUMOS 2014, we elevated action recognition to a more practical level by introducing temporally untrimmed videos. These also include `background videos' which share similar scenes and backgrounds as action videos, but are devoid of the specific actions. The three editions of the challenge organized in 2013--2015 have made THUMOS a common benchmark for action classification and detection and the annual challenge is widely attended by teams from around the world. In this paper we describe the THUMOS benchmark in detail and give an overview of data collection and annotation procedures. We present the evaluation protocols used to quantify results in the two THUMOS tasks of action classification and temporal detection. We also present results of submissions to the THUMOS 2015 challenge and review the participating approaches. Additionally, we include a comprehensive empirical study evaluating the differences in action recognition between trimmed and untrimmed videos, and how well methods trained on trimmed videos generalize to untrimmed videos. We conclude by proposing several directions and improvements for future THUMOS challenges.