Classification of Foot Thermograms using Texture Features and Support Vector Machine

Foot ulcers in Diabetic Mellitus patients are a serious concern as they could lead to amputation of the foot if not detected at early stages. Routine monitoring of the foot using reliable diagnostic tools is essential to assess the risk of developing the ulcer. As ulcers exhibit high vascularity in ulcer-affected areas in the foot, this temperature can be captured and analyzed using image processing techniques. In this paper, 50-foot thermograms are considered for the study. The grey level texture features are calculated in the form of a matrix from the intensity values of normal and diabetic foot thermogram, and they are given as input vectors to the support vector machine classifier. The study performed well in classification and resulted in 96.42 % of accuracy.