Malaria Detection from Blood Smears Using Deep Learning Frameworks
Y. C. A. Reddy Padmanabha, Koushik Reddy Anumula, Joshua Murray · 2026
Malaria is a mosquito-transmitted mosquito-like disease that still kills thousands of people annually. According to increasing deaths, detection is low and laboratory workers are not experienced and sophisticated medical diagnostic equipment is also not available. Recently, research has focused on microscopic blood smear examination of malaria-infected RBCs with DL models as a promising point of care approach. In the present study, the Malaria dataset was used to classify and detect malaria parasites with Hybrid Deep Learning Models. Classification models include ”VGG19, ResNet50, CNN, and advanced RNN” combinations like ”LSTM+LSTM, GRU+GRU, GRU+LSTM, LSTM+GRU, GRU + BiLSTM, LSTM + BiLSTM, and BiLSTM + BiLSTM, alongside Xception, NasNetMobile, and an ensemble model of Xception and NasNetMobile”. For robust detection, the YOLO family, including ’YOLOV5x6, YOLOV5s6, YOLOV8n, and YOLOV9n’, is used to identify abnormalities. The outcome shows that the ensemble Xception + NasNetMobile gave better accuracy, while YOLOV9n proved more effective for malaria parasite detection. Therefore, it shows promising advancement in the area of automated malaria identification and also promotes reliable development of diagnostic tools for malaria control.