Pre-trained Deep Convolutional Neural Network for Detecting Malaria on the Human Blood Smear Images

Abstract

Malaria is a disease caused by the Plasmodium falciparum parasite carried by female Anopheles mosquitoes. This disease is still a severe threat in eastern Indonesia which is an endemic area of Malaria. A data-driven computer-aided diagnostic approach can be an innovative solution. From the experiment results using the Pre-trained Deep Convolutional Neural Network algorithm that was trained with the transfer learning method, the GoogLeNet model was able to achieve a detection accuracy of 93.89%. In comparison, the ShuffleNet V2 model gained 95.20% accuracy with training times three times faster than GoogLeNet.