Development of Computer Assisted Detection Technique for Malaria Parasite using Blood Smear Images

Abstract

Malaria has been the cause of millions of deaths worldwide. According to WHO, the number of deaths due to the disease was 6,19,000 in the year 2021. Given that plasmodia parasites spread to humans through mosquito bites, the matter of concern in cases of malaria is the high mortality rate. The disease causes a highshivering fever with chills accompanied by headaches that take a huge toll on the human body. The elderly and children have a high mortality rate. One of the major reasons for the high mortality rate is the presence of parasites in human blood, which can only be detected 7 days after being bitten by malaria. Antimalarial medication can only be started after confirming the presence of parasites. At times, this delay in treatment proves to be fatal. Traditionally, the detection was done using the blood smear test, in which a sample of blood was analyzed under the microscope. Computer-aided techniques have been able to detect the presence of parasites successfully, which has led to fewer human errors. The proposed study has been undertaken to identify parasites in microscopic blood smear images using computer-assisted techniques. Deep learning techniques can distinguish infected blood cells from uninfected blood cells. Convolutional neural network (CNN) algorithms built on deep learning has greatly improved the precision of parasite detection. The proposed study outperformed earlier approaches with an accuracy rate of 97.24% and also demonstrated an improvement in other crucial factors like the F score and precision, among others. Improved accuracy in the present work will enhance the detection rate of malaria, which may lead to a lower mortality rate because of the deadly disease. The development of a computer-assisted detection technique for malaria parasite using blood smear images holds immense significance for the community and the healthcare industry, as it provides a highly accurate and efficient method for detecting the deadly disease. Future work on improving the accuracy by adding or eliminating