| Summary: | 碩士 === 中原大學 === 醫學工程學系 === 88 === There are four major ways of diagnosis in Chinese medicine practice, tongue inspection is one of the important roles in inspection diagnosis. It is well known that the different sections of tongue are highly correlated to the human organ structure functions. In this research, a complete tongue viewing system was established by using image processing technology. This system could be used for analyzing and finding the difference of tongue characters between healthy persons and Myoma Uteri and Apoplexy and Infecundity patients.
This system was set on a moveable consulting chassis, mainly including a personal computer, a frame grabber, color CCD cameras, and the light equipment. The software which combined with image processing programs and application programs of tongue-characteristic database was developed by using Delphi 5.0. The analyzed results could be saved to tongue-characteristic database for convenience of the preservation and management of patient’s history. The data from database could be use later for statistics and neural network training, then used ANOVA for diseases recognition.
After the system was completely constructed, the difference of tongue characters was generalized by sampling 33 healthy persons and 31 Myoma Uteri patients, 32 Infecundity patients, 31 Apoplexy patients, 55 Upper GI disorder patients. The results showed that the percentage of light red tongue of Myoma Uteri patient was smaller than normal’s, and significant difference between the right and left of tongue temperature compared to normal’s. The significant difference of that Infecundity patient’s tongue area was larger than normal’s, and tongue temperature was smaller than normal’s. The coating percentage Apoplexy patients was significantly larger than normal’s, and tip of tongue temperature was smaller than normal’s. The coating of middle tongue of upper GI disorder patient was larger than normal’s, and middle of tongue temperature was smaller than normal’s. According to tongue characters, the system recognition rate for identifying the normal group, Myoma Uteri, Infecundity, Apoplexy, Upper GI disorder patients groups approached 93%,95%,86% and 85%,respectively. However the recognition rate for training with neural network of the above four groups approached 100%,98%,86% and 93%,respectively. It used ANOVA for diseases recognition, normal group and Upper GI disorder patients groups and Apoplexy were classfied according to ANOVA, however Myoma Uteri group and Infecundity group were not classfied according to ANOVA.
The principles of traditional Chinese medicine diagnosis were based on four diagnostic processes which were viewing, learning, inquiring, and examining. However our developed tongue viewing system and infrared system were only provided tongue diagnostic information. Therefore, the system could be integrated with other Chinese diagnostic systems in the future. It would be practical to help doctors to diagnose illness more accurately.
|
|---|
