DetailVerifyBench: A Benchmark for Dense Hallucination Localization in Long Image Captions
This addresses the need for fine-grained evaluation in long image captioning, which is incremental as it builds on existing benchmarks by adding granularity and diversity.
The paper tackles the problem of detecting and localizing hallucinations in long image captions by introducing DetailVerifyBench, a benchmark with 1,000 images across five domains, featuring captions averaging over 200 words and token-level annotations for multiple hallucination types.
Accurately detecting and localizing hallucinations is a critical task for ensuring high reliability of image captions. In the era of Multimodal Large Language Models (MLLMs), captions have evolved from brief sentences into comprehensive narratives, often spanning hundreds of words. This shift exponentially increases the challenge: models must now pinpoint specific erroneous spans or words within extensive contexts, rather than merely flag response-level inconsistencies. However, existing benchmarks lack the fine granularity and domain diversity required to evaluate this capability. To bridge this gap, we introduce DetailVerifyBench, a rigorous benchmark comprising 1,000 high-quality images across five distinct domains. With an average caption length of over 200 words and dense, token-level annotations of multiple hallucination types, it stands as the most challenging benchmark for precise hallucination localization in the field of long image captioning to date. Our benchmark is available at https://zyx-hhnkh.github.io/DetailVerifyBench/.