Detection of regions with suspected Chronic Venous Insufficiency in thermograms of lower limbs for Mexican Patients

Chronic venous insufficiency (CVI) is a health issue affecting millions of people worldwide, making early detection crucial for preventing complications. This paper proposes a modification in an existing methodology, based on growing regions, to segment high-temperature regions in thermograms of individuals with CVI in their lower limbs. Our proposal, based on a modification of the local temperature variation, was tested using 20 thermal images of Mexican patients diagnosed with CVI. For the experiments reported here, a precision of 0.70, recall of 0.85, F-measure of 0.75 and Jaccard coefficient of 0.61 were obtained. Performance was measured with respect to coincidences with manual annotations. These preliminary results show good potential for detecting areas associated with CVI using thermography. Further experiments using a larger database are needed, in order to assess the generalization of the methodology with respect to a clinical ground truth, as well as to propose ways to enhance the performance of the proposed model.