Técnicas De Inteligencia Artificial Aplicadas A Pruebas Capilaroscopicas Para La Detección De Enfermedades Autoinmunes Que Comprometen La Circulación Sanguínea

Abstract

The main purpose of this paper is the development of an artificial intelligence model for the automatic classification of images, in order to optimize the detection of pathologies through capillaroscopy tests of the nail fold, this technique allows obtaining images of the morphology of the capillaries in the proximal nail fold of the hands. We used a database that consists of 300 images of capillaries corresponding to the nail fold. These images were labeled as healthy or diseased subject depending on the patterns of the capillaries. The method used to classify the images into two classes was transfer learning from a MobileNet V2 base model. The results show that the network is capable of detecting the presence of pathological patterns in the capillaries with a precision of 96.667%.