Li Qingli, Liu Zhi
School of Information Science, East China Normal University, Shanghai, China.
Comput Med Imaging Graph. 2009 Apr;33(3):217-21. doi: 10.1016/j.compmedimag.2008.12.004. Epub 2009 Jan 20.
Human tongue is one of the important organs of the body, which carries abound of information of the health status. Among the various information on tongue, color is the most important factor. Most existing methods carry out pixel-wise or RGB color space classification in a tongue image captured with color CCD cameras. However, these conversional methods impede the accurate analysis on the subjects of tongue surface because of the less information of this kind of images. To address problems in RGB images, a pushbroom hyperspectral tongue imager is developed and its spectral response calibration method is discussed. A new approach to analyze tongue color based on spectra with spectral angle mapper is presented. In addition, 200 hyperspectral tongue images from the tongue image database were selected on which the color recognition is performed with the new method. The results of experiment show that the proposed method has good performance in terms of the rates of correctness for color recognition of tongue coatings and substances. The overall rate of correctness for each color category was 85% of tongue substances and 88% of tongue coatings with the new method. In addition, this algorithm can trace out the color distribution on the tongue surface which is very helpful for tongue disease diagnosis. The spectrum of organism can be used to retrieve organism colors more accurately. This new color analysis approach is superior to the traditional method especially in achieving meaningful areas of substances and coatings of tongue.
人类舌头是身体的重要器官之一，承载着大量健康状况信息。在舌头的各种信息中，颜色是最重要的因素。大多数现有方法对彩色CCD相机拍摄的舌头图像进行逐像素或RGB颜色空间分类。然而，由于这类图像信息较少，这些传统方法妨碍了对舌面主体的准确分析。为了解决RGB图像中的问题，开发了一种推扫式高光谱舌成像仪，并讨论了其光谱响应校准方法。提出了一种基于光谱角映射器的光谱分析舌色新方法。此外，从舌图像数据库中选取了200幅高光谱舌图像，并用新方法进行了颜色识别。实验结果表明，该方法在舌苔和舌质颜色识别的正确率方面具有良好的性能。新方法对每种颜色类别的总体正确率为舌质的85%和舌苔的88%。此外，该算法可以追踪舌面的颜色分布，这对舌病诊断非常有帮助。生物体的光谱可用于更准确地检索生物体颜色。这种新的颜色分析方法优于传统方法，尤其是在获取舌质和舌苔的有意义区域方面。
