SCBench: A Sports Commentary Benchmark for Video LLMs

TL;DR

SCBench introduces a sports video commentary benchmark with a new dataset and metric, enabling detailed evaluation of Video LLMs' temporal and visual understanding in complex, real-world sports scenarios.

Contribution

The paper presents a novel sports commentary generation task, a specialized dataset, and a new six-dimensional evaluation metric for Video LLMs, addressing limitations of existing benchmarks.

Findings

InternVL-Chat-2 achieves top performance with a score of 5.44.

The proposed metric and dataset enable more nuanced assessment of models.

Evaluation reveals current models still have significant room for improvement.

Abstract

Recently, significant advances have been made in Video Large Language Models (Video LLMs) in both academia and industry. However, methods to evaluate and benchmark the performance of different Video LLMs, especially their fine-grained, temporal visual capabilities, remain very limited. On one hand, current benchmarks use relatively simple videos (e.g., subtitled movie clips) where the model can understand the entire video by processing just a few frames. On the other hand, their datasets lack diversity in task format, comprising only QA or multi-choice QA, which overlooks the models' capacity for generating in-depth and precise texts. Sports videos, which feature intricate visual information, sequential events, and emotionally charged commentary, present a critical challenge for Video LLMs, making sports commentary an ideal benchmarking task. Inspired by these challenges, we propose a novel task: sports video commentary generation, developed $\textbf{SCBench}$ for Video LLMs. To construct such a benchmark, we introduce (1) $\textbf{SCORES}$, a six-dimensional metric specifically designed for our task, upon which we propose a GPT-based evaluation method, and (2) $\textbf{CommentarySet}$, a dataset consisting of 5,775 annotated video clips and ground-truth labels tailored to our metric. Based on SCBench, we conduct comprehensive evaluations on multiple Video LLMs (e.g. VILA, Video-LLaVA, etc.) and chain-of-thought baseline methods. Our results found that InternVL-Chat-2 achieves the best performance with 5.44, surpassing the second-best by 1.04. Our work provides a fresh perspective for future research, aiming to enhance models' overall capabilities in complex visual understanding tasks. Our dataset will be released soon.