Software development for digital Chinese medicine tongue diagnosis

• Main Author: Heng, Thong Wah
• Other Authors: Zheng Jianmin
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering::Software
• Online Access: http://hdl.handle.net/10356/59924
• Institution: Nanyang Technological University

Tongue Diagnosis in Traditional Chinese Medicine (TCM) has been around for centuries, and these diagnosis are still practiced to determine the current state of health. Presently, there are sophisticated methodologies and equipment used to perform a detailed scan of the tongue located at medical facilities. However, a quick health diagnostic software that is easily available to users with access to computing systems is currently unavailable. Given today’s hectic working habits, a brief diagnosis on a person’s tongue using a camera and a computing device would provide information on the person’s health status, without a need to be physically present at medical facilities if the person deems that it is not a serious health matter. Previous researches focuses on the variation of tongue analysis techniques and the equipment used to perform diagnosis. These equipment are usually high-edge devices and require an additional camera to be attached in order to perform a diagnosis. Yet, no known convenient applications have been developed to perform a brief diagnosis at a user’s convenience. As such the basis of this project seeks to develop a simple software that will allow users to gain brief information on the current state of their health. The development of this software provides a framework for image analysis and supports the replacement of image comparison algorithms if better algorithms are developed. The iterative model has been used for this project to cater to the type of project. Various methodologies have been identified to suit the image comparison algorithm. One such method includes image histogram comparison to filter out the top three results, followed by having a pixel by pixel comparison to finalize the most similar image. Another faster method is to use a naïve simple comparator, to match the most similar image by transforming the image. However, both methods are unable to yield an efficient analysis timing of about one to two seconds and thus have to be improved on. To conclude, this software aims to provide a quick health diagnostic tool on the human tongue through the capturing and comparison of images. If a more efficient image comparison algorithm is found, it could replace the current algorithm and improve efficiency of the time to display the results.