Intelligent Network Application in Computer-aided Diagnosis

Abstract

Malaria is an infectious disease caused by plasmodium parasites that can be propagated through the bite of female mosquitos. According to WHO's latest World malaria report, an estimated malaria death of 435,000 occurs from 2015 to 2017. Microscopy examination, including stained thin and thick blood smears, is the gold standard for malaria diagnosis. Thick blood smears are used to detect the presence of malaria parasites, and thin blood smears are used to differentiate parasite species. Microscopy examination is of low cost and but is time-consuming and error-prone. Therefore, automatic parasite detection with high accuracy is of important clinical values. To this end, this paper proposes an automatic parasite detection algorithm based on Faster R-CNN, which can automatically detect small objects of malaria parasites. Based on public dataset, we compare our method with ERT and CNN in detection precision. Experimental results show that our method achieves an average precision of 94.61% in the test set.