6:45

MEDSIGHT: Towards Grounded Visual Comprehension in Medical Large Vision-Language Models