EvalCrafter: Benchmarking and Evaluating Large Video Generation Models

TL;DR

This paper introduces EvalCrafter, a comprehensive benchmarking framework for large video generation models that uses a diverse prompt set and multiple metrics, aligning evaluation scores with human preferences.

Contribution

It presents a novel evaluation pipeline with 700 prompts and a human-aligned scoring method, improving assessment accuracy of video generation models.

Findings

The proposed benchmark correlates better with human judgments than traditional metrics.

EvalCrafter enables detailed analysis of visual, content, motion, and alignment qualities.

State-of-the-art models are evaluated comprehensively using the new framework.

Abstract

The vision and language generative models have been overgrown in recent years. For video generation, various open-sourced models and public-available services have been developed to generate high-quality videos. However, these methods often use a few metrics, e.g., FVD or IS, to evaluate the performance. We argue that it is hard to judge the large conditional generative models from the simple metrics since these models are often trained on very large datasets with multi-aspect abilities. Thus, we propose a novel framework and pipeline for exhaustively evaluating the performance of the generated videos. Our approach involves generating a diverse and comprehensive list of 700 prompts for text-to-video generation, which is based on an analysis of real-world user data and generated with the assistance of a large language model. Then, we evaluate the state-of-the-art video generative models on our carefully designed benchmark, in terms of visual qualities, content qualities, motion qualities, and text-video alignment with 17 well-selected objective metrics. To obtain the final leaderboard of the models, we further fit a series of coefficients to align the objective metrics to the users' opinions. Based on the proposed human alignment method, our final score shows a higher correlation than simply averaging the metrics, showing the effectiveness of the proposed evaluation method.