Enhancing Malaria Detection From Blood Smear Images Using Hybrid Deep Learning Models
C Nithisha, K Kushitha, Kuppala Revanth, Sura Veera Surya Narayana Reddy, Alahari Pavana Sree, Reddyvari Srnivas Teja · 2025
Effective treatment of malaria requires rapid and accurate diagnosis, which is still a significant global health problem. The deficiencies of conventional methods of diagnosis lie in their reliance on manual knowledge, emphasizing the need for automated solutions. This work presents a deep learning-based approach that leverages CNNs to improve computing efficiency, reduce errors, and enhance classification accuracy. Our proposed models deliver excellent results; the CNN model, for instance, surpasses several traditional approaches with an impressive accuracy of 99.74%. The ResNet model achieves a high accuracy of 98.76% and is therefore useful for feature extraction as well as classification. The VGGNet model performs well at a 94.55% accuracy. These results indicate that our models are well-equipped to supply reliable and truthful malaria detection. The strong performance of our models along with minimized computing time supports the real-time deployment of models at point-of-care settings in resource-constrained settings. It has been well proven that such research is bringing a step ahead to address the crucial global health problem by unveiling the possibility for efficient, precise, and scalable malaria diagnosis.