ViBe: A Text-to-Video Benchmark for Evaluating Hallucination in Large Multimodal Models

TL;DR

This paper introduces ViBe, a large-scale benchmark dataset of hallucinated videos from Text-to-Video models, to evaluate and improve the detection of AI-generated inconsistencies in video outputs.

Contribution

The paper presents a new dataset and classification framework for identifying hallucinations in T2V models, highlighting the challenge of automated hallucination detection.

Findings

Established baseline classification performance with TimeSFormer + CNN ensemble.

Identified five major hallucination types in T2V outputs.

Demonstrated modest accuracy of current detection methods, emphasizing need for improvement.

Abstract

Recent advances in Large Multimodal Models (LMMs) have expanded their capabilities to video understanding, with Text-to-Video (T2V) models excelling in generating videos from textual prompts. However, they still frequently produce hallucinated content, revealing AI-generated inconsistencies. We introduce ViBe (https://vibe-t2v-bench.github.io/): a large-scale dataset of hallucinated videos from open-source T2V models. We identify five major hallucination types: Vanishing Subject, Omission Error, Numeric Variability, Subject Dysmorphia, and Visual Incongruity. Using ten T2V models, we generated and manually annotated 3,782 videos from 837 diverse MS COCO captions. Our proposed benchmark includes a dataset of hallucinated videos and a classification framework using video embeddings. ViBe serves as a critical resource for evaluating T2V reliability and advancing hallucination detection. We establish classification as a baseline, with the TimeSFormer + CNN ensemble achieving the best performance (0.345 accuracy, 0.342 F1 score). While initial baselines proposed achieve modest accuracy, this highlights the difficulty of automated hallucination detection and the need for improved methods. Our research aims to drive the development of more robust T2V models and evaluate their outputs based on user preferences.