Publication

Abstract : The identification and classification of different kinds of parasite eggs in microscopic samples represent a critical challenge in the field of Soil-transmitted helminth infection diagnosis. Traditional methods are often labor-intensive and time-consuming. The emergence of deep learning models has shown promising results in automating this process by extracting intricate features from complex images. This study aims to develop an automated system for accurately classifying parasite egg types in microscopic images by leveraging the ability of squeeze excitation layers to learn the global information from the input. The proposed system employs features extracted by ResNet50 and ResNet101 with Squeeze Excitation (SE) layers for analysis. The extracted features are then input into a Support Vector Classifier. The study systematically evaluates the features extracted from ResNet50+SE and ResNet101+SE. Results from the evaluation demonstrate the efficacy of the ResNet50+SE in accurately classifying parasite egg types in microscopic images with an accuracy of 0.94. The study provides valuable insights into the choice of squeeze-excitation block added Resnet in the context of parasitology, contributing to the advancement of automated medical image analysis. The findings hold great potential for improving diagnostic processes and supporting epidemiological studies through efficient and accurate parasite detection.