New framework enhances explainability for AI in retinal disease detection

Researchers have developed CounterFundus, a novel framework for explaining deep learning models used in retinal disease classification. This framework utilizes CycleGAN to generate counterfactual explanations, translating pathological images into estimated healthy counterparts. A new metric, the Counterfactual-Classifier Alignment Score (CCAS), quantifies the spatial agreement between these counterfactuals and the classifier's attention, demonstrating consistency with clinically relevant retinal evidence. Furthermore, CCAS-filtered counterfactual augmentation has been shown to improve classification performance in fundus images. AI

IMPACT Enhances trust and clinical adoption of AI in medical imaging by providing interpretable disease localization.

RANK_REASON The cluster contains a research paper detailing a new framework and metric for explainable AI in a specific medical domain. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• Counterfactual-Classifier Alignment Score
• CounterFundus
• CycleGAN
• EfficientNet-B5
• EigenCAM
• Hugging Face
• Kritanu Chattopadhyay
• retinal disease classification

• paper
• product
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →