Research And Development Of Standard Tongue Images Acquisition Device

With the trend of information technology applied to Traditional Chinese Medicine(TCM)digitalization,it is more and more important to describe TCM objectively and quantitatively.As the head of TCM,traditional tongue diagnosis is carried out by the method that doctors observe the tongue and diagnose disease via rich experience.This causes several problems such as weak objectivity,little quantitative and bad sustainability.Nowadays TCM tongue diagnosis has some difficulties in TCM digitalization.It is not consistent in color of digital tongue images captured by same device at different time or different devices.Therefore this results in the fact that tongue images captured from different platforms cannot be analyzed.What's more,tongue images database cannot be compatible,etc.In order to solve these problems,we optimize and designs tongue image acquisition device and has process about tongue image color through experimental research in this paper.In terms of the problem of color inconsistency of tongue image in digitalization process,we propose the method about tongue image acquisition device optimization,design,and color consistency.In this thesis,imaging camera and lighting condition in tongue image acquisition device are analyzed and compared in detail.The imaging quality of camera is evaluated by three criterion-peak signal to noise ratio(PSNR),blur coefficient(Kblur),quality factor(Q);In the aspect of lighting condition,how the length of LED lighting affects the tongue image color is conducted and compared.What's more,LED lighting performance in different lighting condition has been tested.In this thesis,we design a new portable tongue image acquisition device,which is standardized according to experiments.Affected by the physical characterization of device,tongue images captured by standardized different devices exist color difference.For the high precision requirement of tongue image,in the process of tongue image preprocessing,we have a thorough comparison between polynomial calibration algorithm and support vector regression algorithm.What's more,support vector regression algorithm is improved.And performance evaluation about tongue image captured in different environment is implemented according to the criterion of color consistency we made.The algorithm we proposed has a strong ability to calibrate tongue images acquired under different conditions..In order to improve the color consistency of tongue image,we design a new colorchecker used for color calibration of tongue image.By using CIELAB1976 and CIEDE2000 these two formulas,color difference is calculated.Compared with the Munsell colorchecker in original system,color precision is highly increased by using tongue colorchecker.In the meantime,The requirement of color consistency of tongue image is achieved.Experiments indicate that optimized and designed tongue image acquisition device and color calibration can achieve the goal of better color consistency.