T2MBench: A Benchmark for Out-of-Distribution Text-to-Motion Generation

TL;DR

This paper introduces T2MBench, a comprehensive benchmark for evaluating text-to-motion models on out-of-distribution inputs, revealing current models' limitations and guiding future improvements.

Contribution

It presents a new OOD benchmark with a large prompt dataset and a unified evaluation framework for assessing text-to-motion models under complex, out-of-distribution conditions.

Findings

Models excel in semantic alignment but struggle with fine-grained accuracy.

Most models have limited generalization in OOD scenarios.

The benchmark provides practical insights for improving text-to-motion models.

Abstract

Most existing evaluations of text-to-motion generation focus on in-distribution textual inputs and a limited set of evaluation criteria, which restricts their ability to systematically assess model generalization and motion generation capabilities under complex out-of-distribution (OOD) textual conditions. To address this limitation, we propose a benchmark specifically designed for OOD text-to-motion evaluation, which includes a comprehensive analysis of 14 representative baseline models and the two datasets derived from evaluation results. Specifically, we construct an OOD prompt dataset consisting of 1,025 textual descriptions. Based on this prompt dataset, we introduce a unified evaluation framework that integrates LLM-based Evaluation, Multi-factor Motion evaluation, and Fine-grained Accuracy Evaluation. Our experimental results reveal that while different baseline models demonstrate strengths in areas such as text-to-motion semantic alignment, motion generalizability, and physical quality, most models struggle to achieve strong performance with Fine-grained Accuracy Evaluation. These findings highlight the limitations of existing methods in OOD scenarios and offer practical guidance for the design and evaluation of future production-level text-to-motion models.