Automatic Identification of Human Parasite Eggs Based on Multitexton Histogram Retrieving the Relationships between Textons
Roxana Flores-Quispe, Yuber Velazco-Paredes, Raquel E. Patiño-Escarcina, César Beltrán · 2014
In order to identify the parasitic diseases, this paper propose the automatic identification of Human Parasite Eggs to eight different species : Ascaris, Uncinarias, Trichuris, Dyphillobothrium-Pacificum, Taenia-Solium, Fasciola Hepetica and Enterobius-Vermicularis from their microscopic images based on Multitexton Histogram - MTH using new structures of textons. This proposed system includes two stages. In first stage, a feature extraction mechanism that is based on MTH descriptor retrieving the relationships between textons. In second stage, an CBIR system has been implemented in orden to detect their correct species of helminths. Finally, simulation results show overall success rates of 94,78% in the detection.