Design And Implementation Of Tongue Image Data Processing Functions

As the oldest traditional medicine, Traditional Chinese Medicine (TCM) gradually formed an unique theoretical system and treatment technology in its long course of development. TCM has made indelible contributions to the health of the Chinese people. Due to its typical oriental style and remarkable curative effect, TCM is going to play an increasingly important role for the21st century human health. Tongue Diagnosis is an important part of TCM. It obtains information through observing the tongues’ characteristics of patients to get their state of healthiness, and then works out diagnosis and state of an illness. It is one of the most valuable and widely used diagnostic methods in TCM.This dissertation introduces the Relation Database Management System (RDBMS) and regular database developing technology. After analyzing the practical needs of Tongue Database System, the abstraction of data and the E-R graph are made out, and the change from E-R model to Relation Model and Data Model is completed. These are all for the design of Tongue Database System. Tongue Database System not only provides store and search functions for the original tongue images but also stores the feature set which is used in the process of feature extraction and matching and expert knowledge which is used in the process of tongue image classification. Tongue Database System is set up in SQL Server RDBMS and adopts Client/Server mode. In order to keep consistent with the Automatic Chinese Medicine Tongue Diagnosis System (ACMTDS) developing environment, we use Visual C++as the developing tool. After comparing between several developing techniques and doing a lot of experiments we choose ODBC API technique to develop the Tongue Database System. Although it increases the coding work, the interface is much more flexible. It can meet the need in speed and performance and support the ACMTDS very well.This dissertation also explains how to use image-analyzing methods to recognize the texture of tongue image. On the basis of a lot of experiments, we decide to use the method of Gray-level Co-occurrence Matrix (GLCM) statistics to extract the features from the B color plane instead of luminosity to classify the tongue image texture and get a good result. At last a combined method of GLCM and stable moment invariant is proposed, using stable moment invariant of the GLCM as the texture describer, satisfied results are obtained.