Application of an automatic ulcer segmentation algorithm

Legs and feet ulcers are a common and significant public and economic health problem, especially due to the increase in the aging population. These type of wounds usually require a long-term clinical treatment and follow-up. The clinicians perform the diagnosis of skin ulcers on the visual assessment of pathological skin and this make the evaluation highly dependent on the observer and on his visual perception. It is also difficult to have an objective evaluation of the size evolution of the wound to build a patient history. In the last years, many approaches have been proposed to provide a quantitative and objective information on the state of the wound. Automatic ulcers detection and tissue analysis are the main task in this field. In this work, we applied an automatic algorithm for ulcer segmentation to a set of 35 images acquired from 5 different patients with skin ulcers. The segmentation is performed using an automatic processing based on pixels' classification into background or not background classes. Features used to perform the classification are the values of three channels that define each pixel in the RGB color map and in the HSV color map. Images acquired with a smartphone camera were analyzed by a medical expert, which manually segmented the edge of the wound. Then we processed the images with the algorithm and we evaluated the performances. We compared information obtained using the computational approach, with information derived from the manual analysis of the wounds. At first, the method has proved to be applicable in a clinical setting in combination with the use of a low budget device (a smartphone). Furthermore, the results show that the automatic method obtains an objective, repeatable and fast analysis, without loss of significant information.