摘要
皮肤癌的最有效治疗方法是早期诊断加积极有效切除原发灶,对预后和降低死亡率起决定性作用。然而,依靠肉眼对皮肤肿瘤进行诊断,主观性大,即使训练有素的专家其诊断也存在较大的差异。皮肤镜图像计算机辅助诊断系统正是解决这个问题的有效途径,其可以对病变组织自动提取、智能识别,具有定量测量和分析的功能,使诊断更加精确、客观。本文对皮肤镜图像计算机辅助诊断系统的研究现状进行综述,并对皮肤镜图像分析中所涉及的图像质量评价、预处理去噪、皮损分割、特征提取和分类识别等技术进行总结,最后给出未来发展趋势。为此方面的研究人员提供借鉴意义。
Early diagnosisfi）r skin cancer is very important to decrease the mortality. While,it is subjective to diagnose the skin tumor by naked eyes, and even though for experienced experts,the diagnosis results will be very different. Computer-aided diagnosis （CAD） system based on dermoscopy images can extract lesion objects and distinguish the be- nign from the malignant,which is object and accuracy. In this paper,the research sta-tus of CAD system is reviewed firstly;and then the dermoscopy image analysis technology including image quality assessment, preprocessing, lesion segmentation, feature extraction and classification are summarized; finally, the future development trend is given. This paper is a reference for the researchers in dermoscopy image.
出处
《中国医学文摘（皮肤科学）》
2016年第1期45-50,5,共6页
China Medical Abstracts(Dermatology)
关键词
皮肤镜图像
计算机辅助诊断
图像质量评价
图像分割
皮损分类
Dermoscopy image
Computer-aided diagnosis
Image quality assessment
hnage segmentation
Lesion classification
