Do We Need Large VLMs for Spotting Soccer Actions?

TL;DR

This paper explores replacing traditional dense video processing in soccer action spotting with a lightweight, text-based approach using Large Language Models, achieving comparable accuracy without video data processing.

Contribution

It introduces a novel text-based method employing LLMs to identify soccer actions, reducing computational complexity compared to video-based models.

Findings

Language-based approach performs near state-of-the-art in action spotting

Zero video processing required, saving computational resources

Method is scalable and efficient for large-scale match analysis

Abstract

Traditional video-based tasks like soccer action spotting rely heavily on visual inputs, often requiring complex and computationally expensive models to process dense video data. We propose a shift from this video-centric approach to a text-based task, making it lightweight and scalable by utilizing Large Language Models (LLMs) instead of Vision-Language Models (VLMs). We posit that expert commentary, which provides rich descriptions and contextual cues contains sufficient information to reliably spot key actions in a match. To demonstrate this, we employ a system of three LLMs acting as judges specializing in outcome, excitement, and tactics for spotting actions in soccer matches. Our experiments show that this language-centric approach performs effectively in detecting critical match events coming close to state-of-the-art video-based spotters while using zero video processing compute and similar amount of time to process the entire match.