Detection and Identification of Parasite Eggs from Microscopic Images of Fecal Samples

Abstract

Detection and identification of parasite eggs is one of the important tasks in diagnosis of many diseases. Manual detection and identification process of parasite egg is time consuming and prone to error. Automation of this detection and identification process of various parasite eggs can save time and reduce the error in the diagnosis process. In this paper, we proposed a system that automatically detect the parasite eggs present in the microscopic fecal images of pig and identify Ascaris lumbricoides from the detected eggs. In our study, first, different objects are segmented including parasite egg and other non-egg artifacts present in the microscopic images. In the feature extraction stage, we extracted five different types of features from the segmented images. For classification, Artificial Neural Network (ANN) and Multi-class Support Vector Machine (MC-SVM) are used. The experimental result shows about 95% and 93% accuracy rate in identifying Ascaris eggs using MC-SVM and ANN, respectively.