---
title: "Zero-Shot Multimodal Event Spotting in Soccer Broadcasts"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2610.37583v1"
permalink: "https://feed.astralor.com/items/itm_5c8045f423b8c7ab508d7208"
published: "2026-10-27T17:24:35.000Z"
score: 52
channel: "tech"
tags: ["推理效率","多模态","Zero-Shot Multimodal Event","Spotting","Soccer Broadcasts Automatic","Automatic","Each","SoccerNet-trained CALF"]
---

# Zero-Shot Multimodal Event Spotting in Soccer Broadcasts

## 速览

Automatic soccer event spotting supports highlight generation, video indexing, and sports analytics, but existing methods depend on large annotated datasets and soccer-specific training. We present a zero-shot framework…

## 关键点

- Automatic soccer event spotting supports highlight generation, video indexing, and sports analytics, but existing methods depend on large annotated datasets and soccer-specific training.
- We present a zero-shot framework that detects, classifies, and temporally localizes events in full-length soccer broadcasts without task-specific parameter training.
- Each match is segmented into overlapping clips represented by a compact visual storyboard and the aligned commentary transcript, which a general-purpose multimodal model processes jointly.

## 全文

Automatic soccer event spotting supports highlight generation, video indexing, and sports analytics, but existing methods depend on large annotated datasets and soccer-specific training. We present a zero-shot framework that detects, classifies, and temporally localizes events in full-length soccer broadcasts without task-specific parameter training. Each match is segmented into overlapping clips represented by a compact visual storyboard and the aligned commentary transcript, which a general-purpose multimodal model processes jointly. On three manually annotated matches containing 130 events, the best configuration achieves 72.0% F1@5s and outperforms SoccerNet-trained CALF and NetVLAD++ in a cross-dataset setting, while the storyboard representation reduces inference cost by 72% compared with individual frames. We further analyze the effects of model family, input representation, and prompt optimization, and introduce an evidence-grounded excitement-scoring stage that ranks detected events with natural-language explanations for highlight selection.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_5c8045f423b8c7ab508d7208