tween the lesion areas in the images and five clinical indicators, resulting in high accuracy of the classification framework. The proposed method rep