Rethinking Human Evaluation Protocol for Text-to-Video Models: Enhancing Reliability,Reproducibility, and Practicality

TL;DR

This paper introduces the T2VHE protocol, a standardized human evaluation method for text-to-video models that improves reliability, reproducibility, and practicality, while reducing evaluation costs.

Contribution

The paper presents the T2VHE protocol, including metrics, training, and dynamic evaluation, to enhance human assessment of T2V models and address existing evaluation challenges.

Findings

High-quality annotations achieved

Evaluation costs reduced by nearly 50%

Open-source setup facilitates community adoption

Abstract

Recent text-to-video (T2V) technology advancements, as demonstrated by models such as Gen2, Pika, and Sora, have significantly broadened its applicability and popularity. Despite these strides, evaluating these models poses substantial challenges. Primarily, due to the limitations inherent in automatic metrics, manual evaluation is often considered a superior method for assessing T2V generation. However, existing manual evaluation protocols face reproducibility, reliability, and practicality issues. To address these challenges, this paper introduces the Text-to-Video Human Evaluation (T2VHE) protocol, a comprehensive and standardized protocol for T2V models. The T2VHE protocol includes well-defined metrics, thorough annotator training, and an effective dynamic evaluation module. Experimental results demonstrate that this protocol not only ensures high-quality annotations but can also reduce evaluation costs by nearly 50\%. We will open-source the entire setup of the T2VHE protocol, including the complete protocol workflow, the dynamic evaluation component details, and the annotation interface code. This will help communities establish more sophisticated human assessment protocols.