Tecniques d'analisi i classificació d'imatge per la detecció de paràsits de la leishmaniosi
Limon Jacques, Sofia Melissa · UPCommons institutional repository (Universitat Politècnica de Catalunya) · 2017
Leishmaniosis is considered a neglected disease that causes thousands of deaths annually in some countries, specially tropical and subtropical countries. It is caused by protozoa of the genus Leishmania spp., which develop their life cycle between a vertebrate host and an invertebrate vector that transmits the disease. There are various techniques to diagnose leishmaniosis of which manual microscopy is considered to be the standard. There is a need for the development of automatic techniques that are able to detect leishmania parasites in a robust and unsupervised manner. In this document we present and compare two different procedures for automatizing the detection process. The first one uses conventional image processing methods that have been around for some time but are known to be robust in the field. The second one is linked to a more recent and evolving technology called deep learning, and has proven to deliver outstanding results.