中医舌苔厚度的高光谱检测与分析

Hyperspectral detection and analysis of tongue coating thickness in traditional Chinese medicine

关键词:  高光谱  舌诊  舌苔厚度  苔质  舌诊客观化

Key words:  hyperspectral imaging  tongue diagnosis  tongue coating thickness  coating nature  tongue diagnosis objectification

摘要:

[目的] 利用高光谱技术进行中医不同舌苔测量，探索建立中医舌苔厚度的现代技术检测方法。[方法] 对临床门诊患者进行舌象采集，分别拍摄得到同一患者数码舌象与高光谱舌象，共计609例。利用数码舌象确定舌苔类型，依据纳入排除标准筛选出薄白苔组129例、水滑苔组78例、白腻苔组179例、黄腻苔组103例，共计489例为研究对象，按照舌尖、舌中、舌根、舌左、舌右5个不同舌面分区，对比不同类型舌苔之间的高光谱舌象数值及其差异性。[结果] 在舌尖分区，黄腻苔组数值高于薄白苔组且有显著统计学差异（P＜0.01）；在舌中、舌根分区，白腻苔组数值均高于薄白苔组和水滑苔组且有极显著统计学差异（P＜0.001），黄腻苔组数值均高于薄白苔组和水滑苔组且有极显著统计学差异（P＜0.001）；在舌左分区，黄腻苔组数值均高于薄白苔组和水滑苔组且有显著统计学差异（P＜0.01），高于白腻苔组且有统计学差异（P＜0.05）；在舌右分区，黄腻苔组数值高于薄白苔组且有显著统计学差异（P＜0.01），高于水滑苔组且有极其显著统计学差异（P＜0.001）；其他组组间未见明显统计学差异。[结论] 高光谱数值对舌苔厚度的变化呈正相关，提示高光谱技术可以用于舌苔厚度的测定。

Abstract:

[Objective] The paper measures different types of tongue coating in traditional Chinese medicine (TCM) using hyperspectral technology, it explores how to establish a modern technical method to detect the thickness of tongue coating.[Methods] Tongue images were collected from a total of 609 clinical outpatients by taking respectively the digital tongue images and hyperspectral images of the same patient. The digital tongue images were used to determine the type of tongue coating. According to the inclusion and exclusion criteria, a total of 489 cases were selected as the research objects with 129 cases in the thin white coating group, 78 cases in the water sliding coating group, 179 cases in the white greasy coating group, and 103 cases in the yellow greasy coating group. The hyperspectral tongue image values and their differences between different types of tongue coating groups were compared in five different tongue sections: the tip of the tongue, the center of the tongue, the root of the tongue, the left side of the tongue, and the right side of the tongue.[Results] In the tip of the tongue section, the value in the yellow greasy coating group was higher than that in the thin white coating group, and the difference was statistically significant (P < 0.01). In the center and the root of the tongue sections, the value in the white greasy coating group was higher than those in the thin white coating group and the water sliding coating group, and the difference was extremely statistically significant (P < 0.001). The value in the yellow greasy coating group was higher than those in the thin white coating group and the water sliding coating group, and the difference was extremely statistically significant(P < 0.001).In the left side of the tongue section, the value in the yellow greasy coating group was higher than those in the thin white coating group and the water sliding coating group, and the difference was statistically significant (P < 0.01);the same value was also higher than that in the white greasy coating group, and the difference was statistically significant(P < 0.05). In the right side of the tongue section, the value in the yellow greasy coating group was higher than that in the thin white coating group, indicating a statistically significant difference(P < 0.01), and was higher than that in the water sliding coating group, indicating an extremely statistically significant difference (P < 0.001). There were no significant statistical differences among all the other groups.[Conclusion] The hyperspectral value is positively correlated with the change of the tongue coating thickness, suggesting that the hyperspectral technique can be applied to the measurement of the tongue coating thickness.