Erni Rouza, Fatchul Arifin, Suprapto
Soil-Transmitted Helminths (STH) infection remains a significant global health challenge, particularly in regions with inadequate sanitation. While precise early detection is crucial, conventional techniques like microscopy require substantial time and accuracy. This work rigorously examines recent developments in STH detection utilizing machine learning and deep learning techniques. This study pertains to articles published from 2014 until 2024. During the literature selection process utilizing the PRISMA Method, 26 pertinent articles were extracted from the Google Scholar, PubMed, IEEE Xplore, and Scopus databases. The findings indicated that notably Convolutional Neural Networks (CNN) and U-Net algorithms exhibited markedly superior detection accuracy (95-98%) relative to Support Vector Machines (SVM) and Random Forest (RF) (87-92%) respectively. SVM and RF exhibit superior speed but diminished accuracy when applied to tiny datasets. Moreover, there exists significant potential to boost model performance and address data constraints using transfer learning and data augmentation techniques. This study demonstrates that the integration of artificial intelligence with the Internet of Things (IoT) facilitates expedited and more efficient surveillance through real-time detection of STH in endemic regions. Moreover, crowdsourcing and self-supervised learning (SSL) have emerged as methods for the acquisition of annotated data. Significant recent advancements in machine learning and deep learning technologies forecast expedited, more precise, and scalable STH diagnosis in the future. This can be utilized in future global health surveillance, despite limitations such as restricted data and computational resources. © (2024), (Insight Society). All rights reserved.
Doctoral Program of Engineering, Faculty of Engineering, Yogyakarta State University, Yogyakarta, Indonesia; Department of Electronic and Informatics Engineering, Universitas Negeri Yogyakarta, D.I. Yogyakarta, Sleman, Indonesia; Informatics Engineering Study Program, Faculty of Computer Science, Pasir Pengaraian University, Riau, Indonesia
