Action Spotting using Dense Detection Anchors Revisited: Submission to the SoccerNet Challenge 2022

TL;DR

This paper presents a dense detection anchor method for action spotting in soccer videos, achieving state-of-the-art results and first place in the SoccerNet Challenge 2022 by enhancing temporal precision and combining multiple input features.

Contribution

The paper introduces modifications to a dense detection anchor approach, including processing enhancements and late fusion of features, leading to improved challenge performance and new state-of-the-art results.

Findings

Achieved first place in SoccerNet Challenge 2022.

Set a new state-of-the-art on SoccerNet test set.

Improved results through processing modifications and feature fusion.

Abstract

This brief technical report describes our submission to the Action Spotting SoccerNet Challenge 2022. The challenge was part of the CVPR 2022 ActivityNet Workshop. Our submission was based on a recently proposed method which focuses on increasing temporal precision via a densely sampled set of detection anchors. Due to its emphasis on temporal precision, this approach had shown significant improvements in the tight average-mAP metric. Tight average-mAP was used as the evaluation criterion for the challenge, and is defined using small temporal evaluation tolerances, thus being more sensitive to small temporal errors. In order to further improve results, here we introduce small changes in the pre- and post-processing steps, and also combine different input feature types via late fusion. These changes brought improvements that helped us achieve the first place in the challenge and also led to a new state-of-the-art on SoccerNet's test set when using the dataset's standard experimental protocol. This report briefly reviews the action spotting method based on dense detection anchors, then focuses on the modifications introduced for the challenge. We also describe the experimental protocols and training procedures we used, and finally present our results.