Deep Convolutional Neural Network Model for Optical Microscopic Automated Diagnosis of Plasmodium Falciparum Parasites in Sub-Saharan Africa.
Keywords: Deep Convolutional Neural Network, Plasmodium parasite, Diagnosis, Optical microscopy, Digital blood images
Abstract: Deep learning techniques are prevalent in the medical discipline due to their high level of accuracy in disease diagnosis. One such disease is malaria caused by Plasmodium falciparum and transmitted by the female anopheles mosquito. According to the World Health Organisation (WHO), millions of people are infected annually, leading to inevitable deaths in the infected population. Statistical records show that early detection of malaria parasites could prevent deaths and Deep Convolutional Neural Network (CNN) has proved helpful in the early detection of malarial parasites. The human fault is identified to be a major cause of inaccurate diagnostics in the traditional microscopy malaria diagnosis method. Therefore, the method would be more reliable if human expert dependency is restricted or eliminated, and thus, the motivation of this paper. This study presents the application of a Deep Convolutional Neural Network (CNN) to locally generated, low-cost, portable optical microscopic blood films of thin blood smears for automation of P. falciparum parasite detection in the red blood cells. We propose to source automated microscopy blood films (acquired with a digital camera or smartphone) from both private and public clinical parasitology laboratories within Nigeria. We would use the primary data to train our proposed model to be able to predict and classify these blood as infected or uninfected. The work is at the stage of data collection and we hope to commence the research soon. The proposed model is expected to show that early detection of the malaria parasite has the potential to improve patient's survival through the application of deep CNN and as well reduce the involvement of trained human experts in the malaria diagnosis process. Hence, the computational approach to malaria diagnosis helps eliminate the limitations of traditional approaches.
Submission Category: Machine learning algorithms
Submission Number: 60
Loading