Implementation and comparison of two image segmentation techniques on thermal foot images and detection of ulceration using asymmetry

Abstract

Diabetic foot ulceration is the repeated trauma occurring at the region of the feet soles where more pressure is being applied. Ulceration of diabetic foot is very common which frequently lead to the amputation of the foot. Due to ulceration the temperature at these regions increases which is leading to the amputation and results in major wounds at these regions. Conventional methods like Plantar pressure distribution, inspection and palpation do not detect early breakdown in the skin. However temperature distribution measurement can provide information on occurrence of ulceration. Thermal imaging is the technique of determining the temperature distribution measurements on the surface of the skin from the infrared radiation emitted from the surface of the body. These radiations are acquired by the thermal camera and are converted to thermal images. Recent studies confirm that thermal imaging is helpful in the earlier detection of ulceration. The currently available thermography processing techniques are limited to for only general applications. The identification of region of interest with specific anatomic shapes is not possible with these applications. In this proposed work it is aimed to implement and compare two image segmentation techniques for the segmentation of thermal foot images. Also comparison of the two foot sole is done, hence with the help of asymmetry trying to identify ulceration in the subjects at the early stages.