Custom-Built Deep Learning Architecture for Identifying Malaria Parasites in Blood Smear Images
Sujatha Kamepalli, Chennam Anusha · 2023
The Plasmodium parasite, a form of protozoan parasite, invades red blood cells in the human body and is responsible for causing malaria. Malaria is a potentially lethal disease. Malaria diagnosis must be accurate and rapid in order to treat and control the disease efficiently. It is a major public health problem, specifically in tropical areas, and fast and explicit diagnosis is critical for effective treatment. However, due to low contrast, overlapping cells, and other artefacts, visual evaluation of blood smears can be difficult. A Convolutional Neural Networks (CNN) system for accurately detecting parasites associated with malaria in blood smear pictures is proposed in this research. Thousands of photos of diseased and unaffected blood cells made up the data set used in this investigation. Multiple pools and convolutional layers are included in the proposed CNN model, which follows fully linked layers for classification. The model's effectiveness was assessed comprehensively by employing well-established metrics, including The metrics used to evaluate the performance include accuracy and loss etc. This evaluation facilitated a thorough and comprehensive analysis. The findings revealed that the CNN model exhibited exceptional capability in accurately identifying and categorizing malaria cells, achieving an impressive accuracy rate of 0.93. These results clearly highlight the significant promise of this approach for diagnosing and managing malaria.