基于图像处理技术的黑素瘤和克拉克痣的分类研究

Classification of Melanoma and Clark'Nevus Based on Image Processing Techniques

据统计,黑素瘤仅占皮肤癌的11%,死亡率却最高。黑素瘤和良性的克拉克痣的外观特征极为相似,很难区分。计算机图像处理技术的应用可以作为临床医生的辅助工具,提供有助于鉴别皮肤损伤特性的视觉信息。本研究试图从黑素瘤、原位黑素瘤和克拉克痣的皮肤损伤里提取有效的特征,找到合理的分类模型和鉴别皮肤损伤的算法。由于图像采集过程不同的采光会造成的图像颜色差别,因此本文中图像均使用皮肤镜图像。通过分析对比图片中黑素瘤、原位黑素瘤和克拉克痣的特征,应用多种模式分类方案,在实验结果找出了四种有用的分类方法,其中最好的分类方法对黑素瘤和原位黑素瘤有100%辨析率,克拉克痣达到65%的辨析率。

According to the statistics, melanoma accounts for just 11% of all types of Skin cancer, it is responsible for most of the deaths. Melanoma is visually difficult for clinicians to differentiate from Clark nevus lesions which are benign. The application of image processing techniques to these lesions may be useful as an educational tool for teaching physicians to differentiate lesions, as well as for contributing information about the essential optical characteristics for identifying them. This research tried find the most effective features to extract from melanoma, melanoma in situ and Clark nevus lesions, and to find the most effective pattern-classification criteria and algorithms for differentiating those lesions. The color differences between images that occur because of differences in ambient lighting during the photographic process were minimized by the use of dermoscopic images. Differences in skin color between patients was minimized by using normalizing them by means of converting them to relative-color images, and differences in ambient lighting during photography, and the photographic and digitization processes, original color images were normalized by converting them into relative-color images. Tumors in the relative-color images were then segmented out and morphologically filtered. The filtered-tumor features were then extracted and various pattern-classification schemes were applied. Experimentation resulted in four useful pattern classification methods, the best of which was a classification rate of 100% for melanoma and melanoma in situ （grouped） and 65% for Clark nevus.