AutoSoccerPose: Automated 3D posture Analysis of Soccer Shot Movements

TL;DR

AutoSoccerPose introduces a semi-automated pipeline for analyzing 3D soccer shot postures using a new extensive dataset and a non-linear embedding model, advancing understanding of complex sports movements.

Contribution

The paper presents the 3DSP dataset, a novel non-linear pose embedding model (3DSP-GRAE), and a semi-automated pipeline for soccer posture analysis, addressing limitations of previous linear models and datasets.

Findings

Validated on SoccerNet and 3DSP datasets.

Extended posture analysis beyond annotated data.

Provided baseline for future automation efforts.

Abstract

Image understanding is a foundational task in computer vision, with recent applications emerging in soccer posture analysis. However, existing publicly available datasets lack comprehensive information, notably in the form of posture sequences and 2D pose annotations. Moreover, current analysis models often rely on interpretable linear models (e.g., PCA and regression), limiting their capacity to capture non-linear spatiotemporal relationships in complex and diverse scenarios. To address these gaps, we introduce the 3D Shot Posture (3DSP) dataset in soccer broadcast videos, which represents the most extensive sports image dataset with 2D pose annotations to our knowledge. Additionally, we present the 3DSP-GRAE (Graph Recurrent AutoEncoder) model, a non-linear approach for embedding pose sequences. Furthermore, we propose AutoSoccerPose, a pipeline aimed at semi-automating 2D and 3D pose estimation and posture analysis. While achieving full automation proved challenging, we provide a foundational baseline, extending its utility beyond the scope of annotated data. We validate AutoSoccerPose on SoccerNet and 3DSP datasets, and present posture analysis results based on 3DSP. The dataset, code, and models are available at: https://github.com/calvinyeungck/3D-Shot-Posture-Dataset.