Computer aided detection of diabetic foot ulcer using asymmetry analysis of texture and temperature features

Abstract Diabetic foot ulcer is a foremost complication of poorly controlled diabetes mellitus leading to lower extremity amputation. Early identification depends on repeated risk assessment, preferably on a day-to-day basis especially for high-risk patients. The temperature variations in the areas prone to ulcer are higher than non-ulcerous regions of the foot. The ultimate aim of this study was to develop an efficient algorithm for early detection of diabetic foot with infrared thermal images using asymmetry analysis. The left and right foot regions are segmented using region growing method. In normal plantar thermograms, symmetric temperature distributions are observed, whereas, in the case of the diabetic foot, asymmetry was observed between ipsilateral and contralateral regions of the foot. The texture and temperature features are extracted from the 11 regions of interest from the foot and asymmetric analysis was performed for the features extracted from the ipsilateral and contralateral regions of the foot. Support vector machine was used to classify the region of interest into normal and ulcer. The proposed method achieved maximum accuracy of 95.61%, sensitivity of 96.5% and specificity of 92.41%. The performance of the proposed technique shows that it is trustworthy and effective for early identification of pre-signs of ulceration and aids the clinicians in the treatment of the diabetic foot.