DEVIL: Evaluation of Text-to-Video Generation Models:
A Dynamics Perspective

Abstract

Comprehensive and constructive evaluation protocols play an important role in the development of sophisticated text-to-video (T2V) generation models. Existing evaluation protocols primarily focus on temporal consistency and content continuity, yet largely ignoring the dynamics of video content. Dynamics are an essential dimension for measuring the visual vividness and the honesty of video content to text prompts. In this study, we propose an effective evaluation protocol, termed DEVIL, which centers on the dynamics dimension to evaluate T2V models. For this purpose, we establish a new benchmark comprising text prompts that fully reflect multiple dynamics grades, and define a set of dynamics scores corresponding to various temporal granularities to comprehensively evaluate the dynamics of each generated video. Based on the new benchmark and the dynamics scores, we assess T2V models with the design of three metrics: dynamics range, dynamics controllability, and dynamics-based quality. Experiments show that DEVIL achieves a Pearson correlation exceeding 90% with human ratings, demonstrating its potential to advance T2V generation models.

Dynamics score

By altering the temporal granularities, video dynamics can be categorized into three: (i) Inter-frame dynamics, which describes the variations between successive frames. The dynamics score at this level reflects rapid and prominent content variations. (ii) Inter-segment dynamics, which refers to the changes between video segments which contains K video frames. Define on a middle-level, this score captures middle-speed transitions and motion patterns. (iii) Video-level dynamics, which encompasses the overall content diversity and the frequency of changes throughout the video.

Evaluating Generated Videos

Evaluating Real Videos