Classification method and device for infantile hemangioma skin lesion images

The invention provides an infant hemangioma skin lesion image classification method and device, and belongs to the technical field of medical auxiliary diagnosis. The infant hemangioma skin lesion image classification method comprises the following steps: establishing an infant hemangioma skin lesion image data set; extracting a saliency region in a to-be-classified image in the infant hemangioma skin lesion image data set, and positioning the position of a target object in the saliency region; and extracting saliency components in the target object, and endowing each saliency component with different semantics. According to the infant hemangioma skin lesion image classification method and device provided by the invention, the skin lesion image of the infant hemangioma can be classified more finely. 本发明提出一种婴儿血管瘤皮损图像分类方法及装置，属于医学辅助诊断技术领域。所述婴儿血管瘤皮损图像分类方法包括:建立婴儿血管瘤皮损图像数据集；提取所述婴儿血管瘤皮损图像数据集中待分类图像中的显著性区域，并在所述显著性区域中定位到目标对象的位置；提取所述目标对象中的显著性部件，为各所述显著性部件赋予不同的语义。本发明提出的婴儿血管瘤皮损图像分类方法及装置能够对婴儿血管瘤的皮损图像进行更加精细地分类。