Smartphone Application for Automated Malaria Screening

Ian MatthewsNews

A publication from BMC Infectious Diseases in November 2020 details a smartphone application for automated malaria screening. The Android application combines multiple functions, including image acquisition, screening, and management of the acquired data. Used in combination with a microscope adapter, this results in a very affordable design. Android smartphones and microscopes are both common in malaria clinics, and an adapter is usually inexpensive. According to the inventors, the system has great potential to assist with malaria diagnosis in resource-limited areas because of the low-cost design and easy-to-use interface.

Malaria Screener, available on Google Play, is a fast, low-cost smartphone app for malaria screening that offers important functionalities with an intuitive user interface which automatically screens slides and counts infected red blood cells and parasites in thin and thick smear images for P. Falciparum malaria. It also manages the images and metadata generated throughout the screening process, which can be used to further optimize the image analysis model.

Malaria Screener is the first smartphone-based system that can screen thin and thick smears. For a single thin smear image, the goal is to detect the number of infected red blood cells (RBCs) and the total number of RBCs in the image; the goal for a thick smear is to detect the number of parasites and white blood cells (WBCs). The customized camera function includes a Camera object that controls the intrinsic parameters of the camera hardware, and a Camera Preview object that displays the preview image to the user. During a screening session, the user captures a suitable image and saves it in PNG format. The captured image is then passed to the parasite detection module as input.

The app aims to support the daily work of field workers by integrating additional functions apart from the basic slide screening functions, based on computational image analysis and machine learning. As an example, the data management function allows users to enter patient information directly into the app’s database, avoiding separate data management systems is very helpful. The app offers a powerful and efficient tool for field tests and data collection during malaria research efforts, which are usually done through a collaboration between medical imaging research groups and hospitals.

Malaria Screener resolves the effort involved in data processing and formatting by integrating a slide screening module, a database module, and a data upload module into the same app, making slide screening and data collection a streamlined process that generates and delivers ready-to-use data. The app is a step towards automating malaria light microscopy.

The open-source codebase and modular architecture aim to assist other groups new to research and allows it to be adapted by fellow researchers to advance the field. The modular design enables other groups to build on the current implementation, testing their own parasite detection algorithms. The inventors hope that by making Malaria Screener an open-source project, it will provide a platform for the scientific community to collaborate and to advance the automation of malaria diagnosis.

Ian MatthewsSmartphone Application for Automated Malaria Screening