Eyes on the Image: Gaze Supervised Multimodal Learning for Chest X-ray Diagnosis and Report Generation

Researchers have introduced a two-stage multimodal framework named MIMIC-Ey designed to improve how artificial intelligence interprets chest X-rays. By incorporating gaze-supervised learning, the system aims to better align model attention with the specific diagnostic patterns used by radiologists. This approach seeks to address current limitations in automated report generation and the spatial accuracy of medical image analysis.