White Blood Cells Classification Using Deep Learning Technique

Abstract

Collaboration with technology is the next revolution in health care, and they must adapt evolving healthcare technologies to be significant in the next years. Digitization in medicine and health care might help transform unsustainable healthcare systems into sustainable ones, enhance interactions between medical specialists and people, and provide inexpensive, speedier, and more effective illness remedies. The objective of this study is to devise a system that distinguishes the types of white blood cells using precise and programmed analysis, which has been considered an effective way of classifying the 4 different types of WBCs—eosinophil, lymphocyte, monocyte, neutrophil. Convolutional neural network (CNN) models had been used for detection and classification, providing an efficient solution with a good accuracy rate when compared to the existing system. The training was done using a publicly available dataset, and visualization methods were used to map the model accuracy and loss. It was discovered that neural networks can record the colors, and textures of lesions that are specific to their type, which resembles human decision-making. The final model is deployed using Django framework for classification, where the image is uploaded and the application displays the result.