脾虚舌象识别的图像特征提取优化算法研究

发布时间：2018-03-29 14:39

本文选题：舌诊　切入点：齿痕舌　出处：《江西理工大学》2017年硕士论文

【摘要】：舌诊历来就是传统中医四诊中关于望诊‖的一种重要方式,沿用至今都有重大的医学参考价值。大量研究表明,齿痕舌作为一种脾虚症状的异常舌态,其呈现的形体特征,往往与正常舌体有很大区别,而正是这些区别预示着某种病症。传统的舌诊诊断结果,大多数是依赖于中医专家的主观判断,诊断结果必然会受到外界环境的干扰且往往不具有可重复性。本课题—脾虚舌象识别的图像特征提取优化算法研究,就是针对上述情况而设立的。本课题是以齿痕舌象为切入点,通过提出一些改进算法和分析方法来分割出正确的齿痕舌体及对其相关特征进行分析,主要包括利用基于阈值、区域和边缘检测的算法对齿痕舌像进行分割、利用灰度投影法实现舌像粗分割、对比度增强、利用改进的控制标记符分水岭算法实现齿痕舌像较好分割;结合B样条小波变换和Canny边缘检测算子提取齿痕舌体的边缘信息、采取优化Snakes模型实现舌像分割;对舌象形体、直方图、颜色特征进行提取分析,利用圆形度参数及颜色特征对齿痕舌的剧烈程度进行了判别并划分了脾虚症状的严重程度。实验结果表明,通过改进的算法能实现齿痕舌像的较好分割,在舌体分割的基础上,结合舌像特点及中医有关于舌诊的理论,给出了几组常见的齿痕舌特征参数,针对最具有代表齿痕舌意义的圆形度参数及颜色特征进行了定量提取与分析,利用得到的特征参数能够对齿痕舌的剧烈程度进行有效表征及用于划分脾虚症状的严重程度。
[Abstract]:Tongue diagnosis has always been an important way of looking for diagnosis in traditional Chinese medicine, which has great medical reference value up to now. A large number of studies have shown that as an abnormal tongue state of spleen deficiency symptom, tooth mark tongue presents its physical characteristics. It is often quite different from the normal body of the tongue, and it is these differences that indicate a disease. The traditional diagnosis of the tongue depends mostly on the subjective judgment of the TCM expert. The diagnosis results will inevitably be disturbed by the external environment and will not be repeatable. This subject is based on the tongue image of the tooth mark, and proposes some improved algorithms and analysis methods to segment the correct tongue body of the mark and analyze its relevant features, including the use of threshold value. The algorithm of region and edge detection is used to segment the tongue image of the tooth mark, the coarse segmentation of the tongue image is realized by the gray projection method, the contrast is enhanced, and the improved watershed algorithm of the control marker is used to realize the better segmentation of the tongue image of the tooth mark. Combined with B-spline wavelet transform and Canny edge detection operator to extract the edge information of the tongue body, adopt the optimized Snakes model to realize the tongue image segmentation, extract and analyze the shape, histogram and color feature of the tongue image. The roundness parameters and color characteristics are used to distinguish the severity of the tongue with tooth marks and the severity of the symptoms of spleen deficiency is divided. The experimental results show that the improved algorithm can achieve better segmentation of the tongue image of the teeth mark, based on the segmentation of the tongue body. Combined with the characteristics of tongue image and the theory of tongue diagnosis in traditional Chinese medicine, some common tongue characteristic parameters of tooth mark are given, and quantitative extraction and analysis of roundness parameter and color feature, which are most representative of tooth mark tongue, are carried out. The obtained characteristic parameters can be used to characterize the severity of tongue marks and to distinguish the severity of spleen deficiency symptoms.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R241;TP391.41

【参考文献】