MedHEval: Benchmarking Hallucinations and Mitigation Strategies in Medical Large Vision-Language Models
This addresses the need for reliable medical AI by providing a standardized evaluation framework, though it is incremental as it builds on existing hallucination research.
The paper tackled the problem of hallucinations in Medical Large Vision-Language Models by introducing MedHEval, a benchmark that categorizes hallucinations into three causes and evaluates mitigation strategies, finding that current models struggle with knowledge- and context-based errors and mitigation methods are limited.
Large Vision Language Models (LVLMs) are becoming increasingly important in the medical domain, yet Medical LVLMs (Med-LVLMs) frequently generate hallucinations due to limited expertise and the complexity of medical applications. Existing benchmarks fail to effectively evaluate hallucinations based on their underlying causes and lack assessments of mitigation strategies. To address this gap, we introduce MedHEval, a novel benchmark that systematically evaluates hallucinations and mitigation strategies in Med-LVLMs by categorizing them into three underlying causes: visual misinterpretation, knowledge deficiency, and context misalignment. We construct a diverse set of close- and open-ended medical VQA datasets with comprehensive evaluation metrics to assess these hallucination types. We conduct extensive experiments across 11 popular (Med)-LVLMs and evaluate 7 state-of-the-art hallucination mitigation techniques. Results reveal that Med-LVLMs struggle with hallucinations arising from different causes while existing mitigation methods show limited effectiveness, especially for knowledge- and context-based errors. These findings underscore the need for improved alignment training and specialized mitigation strategies to enhance Med-LVLMs' reliability. MedHEval establishes a standardized framework for evaluating and mitigating medical hallucinations, guiding the development of more trustworthy Med-LVLMs.