Temporally Precise Action Spotting in Soccer Videos Using Dense Detection Anchors

TL;DR

This paper introduces a dense detection anchor-based model for precise action spotting in soccer videos, leveraging large temporal contexts and advanced training techniques to improve localization accuracy.

Contribution

The paper proposes a novel dense detection anchor approach with two architectures and best practices, achieving state-of-the-art results in soccer action spotting.

Findings

State-of-the-art accuracy on SoccerNet-v2 dataset

Predicting temporal displacements improves localization

Training with SAM and mixup enhances performance

Abstract

We present a model for temporally precise action spotting in videos, which uses a dense set of detection anchors, predicting a detection confidence and corresponding fine-grained temporal displacement for each anchor. We experiment with two trunk architectures, both of which are able to incorporate large temporal contexts while preserving the smaller-scale features required for precise localization: a one-dimensional version of a u-net, and a Transformer encoder (TE). We also suggest best practices for training models of this kind, by applying Sharpness-Aware Minimization (SAM) and mixup data augmentation. We achieve a new state-of-the-art on SoccerNet-v2, the largest soccer video dataset of its kind, with marked improvements in temporal localization. Additionally, our ablations show: the importance of predicting the temporal displacements; the trade-offs between the u-net and TE trunks; and the benefits of training with SAM and mixup.