EasyLens: A Training-Free Plug-and-Play Subtle-Lesion Representation Amplifier for Medical Vision-Language Models

2026-06-04

AI Summary

Current medical vision-language models (VLMs) exhibit insufficient sensitivity to subtle lesions due to their sparse, low-contrast visual cues, which are often diluted by complex anatomical backgrounds. To address this limitation, this work proposes EasyLens—a training-free, plug-and-play module that enhances lesion perception without fine-tuning the frozen VLM. EasyLens constructs a pathology-anatomy prototype space (EasyBank), employs counterfactual prototype reasoning (EasyTag) to identify lesion-relevant image patches, and applies morphology-guided residual enhancement (EasyAmplifier) to amplify their representations. This approach achieves state-of-the-art performance across multiple medical imaging datasets and backbone architectures, significantly outperforming existing encoder-enhancement baselines while offering strong generalizability and resistance to overfitting.

Abstract

Medical vision-language models (VLMs) have shown increasing potential for clinical image interpretation, including lesion detection and report generation. However, their practical utility remains limited by insufficient sensitivity to subtle lesions, whose visual evidence is often sparse, low-contrast, and embedded within complex anatomical context. As local visual tokens are aggregated, these weak lesion cues can become underrepresented in global image representations, making them difficult for medical VLMs to recognize. Existing efforts to improve lesion sensitivity mainly rely on medical-domain vision-encoder pre-training, clinical-term-guided alignment, or trainable pathological representation enhancement. Although effective, these approaches usually require additional training or model-specific adaptation and may overfit to particular disease morphologies, limiting their applicability to frozen medical VLMs. To address these limitations, we propose EasyLens, a training-free plug-and-play subtle-lesion representation amplifier for medical VLMs. EasyLens first constructs EasyBank, a pathology-anatomy prototype space that provides lesion-related prototypes and anatomy-aware normal references for comparing suspicious patches against both pathological and normal anatomical patterns. To avoid blindly amplifying normal tissues, EasyTag selects lesion-relevant patches through counterfactual prototype reasoning. To counteract the dilution of subtle lesion cues in global image representations, EasyAmplifier strengthens the selected lesion-relevant patch representations through morphology-guided residual enhancement, thereby increasing their contribution to the global image embedding. Experiments on multiple medical image datasets and frozen medical VLM backbones show that EasyLens improves subtle-lesion detection and outperforms existing encoder-enhancement baselines.

Problem

• subtle lesions
• medical vision-language models
• lesion sensitivity
• frozen VLMs
• visual representation

Innovation

• training-free
• subtle-lesion amplification
• prototype-based reasoning
• plug-and-play module
• medical vision-language models