Smartphone based image analysis for malaria diagnosis

Project Overview

Automated cell segmentation of malaria blood smear image using convoluted neural networks


Malaria is caused by parasites that are transmitted through the bites of infected mosquitoes. With about 200 million cases worldwide, and about 600,000 deaths per year, malaria is a major burden on global health. The greatest impact of the disease is recorded in sub-Saharan Africa and South East Asia, where a death is recorded every minute from malaria, and where malaria is a leading cause of childhood neuro-disability. Typical symptoms of malaria include fever, fatigue, headaches, and in severe cases seizures, coma, and death. While existing drugs make malaria a curable disease, inadequate diagnostics and emerging drug resistance are major barriers to successful mortality reduction. The development of a fast and reliable diagnostic test is therefore one of the most promising ways of fighting malaria, together with better treatment, development of new malaria vaccines, and mosquito control.

The current standard method for malaria diagnosis in the field is light microscopy of blood films. About 170 million blood films are examined every year for malaria, which involves manual counting of parasites, and white blood cells that indicate the severity of infection. Accurate parasite counts are essential to diagnosing malaria correctly, testing for drug-resistance, measuring drug-effectiveness, and classifying disease severity. However, microscopic diagnostics are not standardized and depend heavily on the experience and skill of the microscopist. It is common for microscopists in low-resource settings to work in isolation, with no rigorous system in place that can ensure the maintenance of their skills and thus diagnostic quality. This leads to incorrect diagnostic decisions in the field. For false negative cases, this means unnecessary use of antibiotics, a second consultation, lost days of work, and in some cases progression into severe malaria. For false positive cases, a misdiagnosis entails unnecessary use of anti-malaria drugs and suffering from their potential side effects, such as nausea, abdominal pain, diarrhea, and sometimes severe complications.

To improve malaria diagnostics, the Lister Hill National Center for Biomedical Communications, an R&D division of the U.S. National Library of Medicine, NIH and Mahidol-Oxford Tropical Medicine Research Unit, are developing a fully automated low-cost system that uses a mobile phone and standard microscope for parasite detection and counting in microscopic images of blood films. Automatic parasite counting has several advantages compared to manual counting: it provides a more reliable and standardized interpretation of blood films; it allows more patients to be served by reducing the workload of the malaria field workers; and it reduces diagnostic costs. To count parasites automatically, the system uses image processing methods to find cells infected with parasites in digitized images of blood films. The system is trained on manually annotated images and machine learning methods then discriminate between infected and uninfected cells, detect the type of parasites that are present, and perform the counting. The system uses a regular smartphone and digital images acquired on standard light microscopy equipment, which makes it well suited for resource-poor settings.

Proposed Work

The goal of this project is to develop the system for real-world use for malaria diagnosis. It will include optimisation of the system at NIH and testing of the system in the field at MORU. This will comprise several stages: 1. Testing and optimisation of the smartphone application interface and performance at NIH; 2. Testing and optimisation of the system for connecting the smartphone to standard light microscopes at NIH and at MORU in Bangkok; 3. Development of a core set of performance metrics for the application; 4. Preliminary field testing of the entire system for malaria diagnosis together with government healthcare workers and National Malaria Control Programme staff in Bangladesh and Thailand; 5. Structured interviews to gather feedback on the system and its potential role in malaria diagnosis in different settings; 6. Formal field trial of the system; 7. Development of a system implementation guidance document for National Malaria Control Programmes.

Training Opportunities

Students will join a dynamic team of image analysis specialists at NLM and epidemiologists, modellers and clinicians at the MORU offices in Bangkok. The student will spend time at field sites in malaria-endemic areas and will interact with government staff. Training will be provided at NIH on basic image analysis and smartphone application development and at MORU on malaria miscroscopy, clinical study methodology, data analysis and research ethics. Students will also have access to epidemiology training resources at MORU, including an on-site library; will join weekly epidemiology team meetings and journal clubs, as well as training sessions on coding and weekly scientific seminars, plus generic skills training for postgraduate students.


Tropical Medicine & Global Health and Clinical Trials & Epidemiology


Project reference number: 1004

Funding and admissions information


Name Department Institution Country Email
Professor Richard J Maude Tropical Medicine Oxford University, Bangkok THA
Stefan Jaeger Lister Hill Institute National Library of Medicine/National Institutes of Health USA

Rajaraman S, Antani SK, Poostchi M, Silamut K, Hossain MA, Maude RJ, Jaeger S, Thoma GR. 2018. Pre-trained convolutional neural networks as feature extractors toward improved malaria parasite detection in thin blood smear images. PeerJ, 6 pp. e4568. Read abstract | Read more

Malaria is a blood disease caused by the parasites transmitted through the bite of female Anopheles mosquito. Microscopists commonly examine thick and thin blood smears to diagnose disease and compute parasitemia. However, their accuracy depends on smear quality and expertise in classifying and counting parasitized and uninfected cells. Such an examination could be arduous for large-scale diagnoses resulting in poor quality. State-of-the-art image-analysis based computer-aided diagnosis (CADx) methods using machine learning (ML) techniques, applied to microscopic images of the smears using hand-engineered features demand expertise in analyzing morphological, textural, and positional variations of the region of interest (ROI). In contrast, Convolutional Neural Networks (CNN), a class of deep learning (DL) models promise highly scalable and superior results with end-to-end feature extraction and classification. Automated malaria screening using DL techniques could, therefore, serve as an effective diagnostic aid. In this study, we evaluate the performance of pre-trained CNN based DL models as feature extractors toward classifying parasitized and uninfected cells to aid in improved disease screening. We experimentally determine the optimal model layers for feature extraction from the underlying data. Statistical validation of the results demonstrates the use of pre-trained CNNs as a promising tool for feature extraction for this purpose. Hide abstract

Poostchi M, Silamut K, Maude RJ, Jaeger S, Thoma G. 2018. Image analysis and machine learning for detecting malaria. Transl Res, 194 pp. 36-55. Read abstract | Read more

Malaria remains a major burden on global health, with roughly 200 million cases worldwide and more than 400,000 deaths per year. Besides biomedical research and political efforts, modern information technology is playing a key role in many attempts at fighting the disease. One of the barriers toward a successful mortality reduction has been inadequate malaria diagnosis in particular. To improve diagnosis, image analysis software and machine learning methods have been used to quantify parasitemia in microscopic blood slides. This article gives an overview of these techniques and discusses the current developments in image analysis and machine learning for microscopic malaria diagnosis. We organize the different approaches published in the literature according to the techniques used for imaging, image preprocessing, parasite detection and cell segmentation, feature computation, and automatic cell classification. Readers will find the different techniques listed in tables, with the relevant articles cited next to them, for both thin and thick blood smear images. We also discussed the latest developments in sections devoted to deep learning and smartphone technology for future malaria diagnosis. Hide abstract