Detection & classification of diabetes type II & diabetes neuropathy using foot images

• Main Author: Chen, Yun Xuan.
• Other Authors: Ng Yin Kwee
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Bioengineering
• Online Access: http://hdl.handle.net/10356/16863
• Institution: Nanyang Technological University

Diabetes is a disease that affects millions of people every year. People with diabetes can develop nerve problems at any time, but the longer a person has diabetes, the greater the threat. In diabetic patients, sensation loss will lead to skin ulceration and thus may result in amputation. Therefore, it is very important to understand and detect the factors that are responsible for plantar ulceration and the measurement being made is necessary to save the foot at risk. Hence, there is a need to detect diabetes at an early stage and start medication early to prevent the onset of diabetic neuropathy. In this work, images of normal, diabetic without neuropathy and diabetic with neuropathy patients are taken and analyzed. The images will be scaled and it will be processed using Continuous Wavelet Transform (CWT). After CWT, as there are many coefficients, hence with the use of matlab, an algorithm of Principle Component Analysis (PCA) is developed. It is used to reduce the number of coefficients so that the data can be analyzed easily. After the reduction of the data with the help of Principle Component Analysis (PCA), the data is being fed into the classifiers, Gaussian Mixture Model (GMM) and Support Vector Machine (SVM). Thereafter, Graphic User Interface (GUI) in Matlab is applied. When an image is selected into the system, it will classify if the patient is normal, diagnosed with Diabetes Type I/II or Diabetic neuropathy. Therefore, this display of classification may help the clinician to effectively identify between normal, early and advanced stages of diabetic neuropathic subjects and also in detecting foot sole areas at risk.