Deep Learning‐Based Image Classifier for Malaria Cell Detection
Alok Negi, Kumar Krishan, Prachi Chauhan · 2021
Malaria is now a lethal, contagious, mosquito-borne critical disease spawn by Anopheles mosquito bitten spread through Plasmodium parasites. Modern information technology, along with biomedical research and political efforts, plays an important role in many attempts to combat the disease. Automating the diagnostic process would allow for precise diagnosis of the infection and therefore holding the potential of providing quality healthcare to resource-scarce areas. Combined to open access resources, artificial intelligence (AI) will boost malaria treatment and a strong mix to improve society. Throughout this chapter, we also explore whether AI can be able to leverage for low-cost, reliable, and precise open sourced deep learning solutions to predict lethal malaria. We outline a few of the findings regarding the reasonably precise malaria-infected cells classification leveraging deep neural convolution (CNN) networks. We will discuss the strategies for assembling a pathologist-cured visual dataset to train deep neural networks, and also some data augmentation approaches used to significantly increase the dimensionality of the data, given the overfitting problem especially in deep CNN training.