CVPR2025

Video-Bench: Human-Aligned Video Generation Benchmark

Hui Han, Siyuan Li, Jiaqi Chen, Yiwen Yuan, Yuling Wu, Yufan Deng, Chak Tou Leong, Hanwen Du, Junchen Fu, Youhua Li, Jie Zhang, Chi Zhang, Li-jia Li, Yongxin Ni

DOI 出版方

摘要

https://github.com/Video-Bench/Video-Bench.git Figure 1. Overview of Video-Bench. Left: We introduce comprehensive evaluation dimensions in two main categories: video-condition alignment (Sec. 3.1.1) and video quality (Sec. 3.1.2). Right: For these two types of dimensions, we analyze their challenges (Sec. 4.1) and corresponding propose the MLLM-based automated video evaluation suite.