A Hybrid Adaptive Object Identification Algorithm for Neglected Tropical Disease Diagnosis

This paper provides an insight into the cross-disciplinary work that combines image processing principles and pathological analysis to diagnose Neglected Tropical Diseases (NTDs). The work described is a novel image processing algorithm that automatically identifies parasitic eggs from Kato-Katz images, in particular, Ascaris and Taenia species. The research outcomes have the potential to be further used for identifying other species that are present in the Kato- Katz images such as Schistosoma Mansoni. The algorithm described in this paper is robust and is able to identify the parasitic eggs at varying levels of magnification and orientation. This algorithm was tested on 4 image sets containing almost 1000 images of Ascaris and Taenia, at lOx and 40x magnification levels.