Malarial Diagnosis with Deep Learning and Image Processing Approaches

Abstract

Malaria is a mosquito-borne disease that has killed an estimated a half-a-million people worldwide since 2000. It may be time consuming and costly to conduct thorough laboratory testing for malaria, and it also requires the skills of trained laboratory personnel. Additionally, human analysis might make mistakes. Integrating denoising and image segmentation techniques with Generative Adversarial Network (GAN) as a data augmentation technique can enhance the performance of diagnosis. Various deep learning models, such as CNN, ResNet50, and VGG19, for recognising the Plasmodium parasite in thick blood smear images have been used. The experimental results indicate that the VGG19 model performed best by achieving 98.46% compared to other approaches. This study demonstrates the potential of artificial intelligence to improve the speed and precision of pathogen detection which is more effective than manual analysis.