Analysis and experimentation of plantar foot segmentation from thermographic digital images for preventive diagnosis of diabetic foot

Plantar foot surface temperature is an important feature in type II diabetes as it is an early sign of foot ulcer. We have reviewed the scientific and technical foundations for the use of digital thermography of the sole of the foot as a prevention procedure to avoid foot ulceration. This project has progressed in the implementation of a new segmentation algorithm to analyze plantar foot temperature and its assessment by qualitative judgment. We have processed thermographic images of the feet. To achieve this we first proceeded to capture images from two soles of feet -surrounded or not by a foam block- and then to segment them using two different methods each of which include the iterative closest point (icp) method: 1) fuzzy clustering modeling (FCM) and 2) growing seeds. Images properly segmented were judged by three observers who have estimated the quality of the contour tracing. Medians and quartile deviations are calculated to estimate the segmentation’s quality. As a result, the quality contour tracing was less suitable when there was not a foam block surrounding the sole of foot while imaging was performed. Likewise, the segmentation using FCM was less suitable than the segmentation by growing seeds. Overall, these results suggest the possibility of avoid using foam block by optimizing the software with growing seed method which increase the comfort of patients and asepsis inside hospitals.