Applications of Signatures Curves to Characterize Melanomas and Moles
Applications of Signatures Curves to Characterize Melanomas and Moles
复制标题
应用特征曲线来表征黑色素瘤和痣
DOI：
--
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
Chehrzad Shakiban
中科院分区：
文献类型：
--
作者：
A. Grim;Chehrzad Shakiban
本文主要研究了利用闭曲线的曲率和曲率对弧长的导数（Sigma ={（kappa（t），kappa_{s}（t））}）构成的欧氏不变曲线（签名曲线）来分析黑色素瘤和痣的轮廓。我们计算皮肤病变轮廓的特征曲线，以检测皮肤病变的不对称性、边界不规则性和直径大小。通过分析60例良性痣和60例恶性黑色素瘤的特征曲线，我们发现良性和恶性病变在其特征曲线上具有不同的全局和局部对称模式。我们还将证明，正常的痣显示出高度的整体对称性，而黑色素瘤表现出多种类型的局部对称性，嵌入其签名曲线。然后，我们将注意力转向ABCD方法的C方面，通过分析黑色素瘤和痣的颜色。最后，我们使用ROC分析，一个关键的统计工具，来分析我们的方法的性能。
In this paper, we focus on the application of an Euclidean invariant curve, called the signature curve, formed by taking curvature and derivative of curvature with respect to arc length of a closed curve, (Sigma = {(kappa (t), kappa _{s}(t))}) to analyze the contour of melanomas and moles. We calculate the signature curves of the contours of the skin lesions to detect asymmetry, boundary irregularity and diameter size of the skin lesions. By analyzing the signature curves of 60 benign moles and 60 melanomas, we show that the benign and malignant lesions have different global and local symmetry patterns in their signature curves. We will also demonstrate that the regular moles show a high degree of global symmetry, whereas melanomas exhibit multiple types of local symmetry that are embedded within their signature curves. We then turn our attention to the C aspect of the ABCD method by analyzing the color of melanomas and moles. Finally, we use ROC Analysis, a key statistical tool, to analyze the performance of our method.