Enhanced Anaemia Detection using Deep Learning Models Integrated into a Web-based Fingernail Imaging Application with Email Authentication
N. Mythili, Anu Varshini S K, Gayathri Devi T · 2025
Anaemia is a known health problem affecting millions of people across the world, especially in substandard areas which cannot afford to use other traditional methods in diagnosing the disease. The presented work outlines the creation of a web-based application for the early and non-invasive diagnosis of anemia based on deep learning models, namely ResNet50, EfficientNetB9, and DenseNet169 using fingernail images. It is intended to be a fast, precise, and convenient diagnostic system which can work on any device that has internet connection. Compared with traditional visible signs of anemia, the application employs fingernail color and surface depiction as an innovative way of diagnosing anemia. In order to protect the user information, procedures such as email authentication and data encryption is used in the system. For the evaluation purpose, all the four parameters of accuracy, precision, recall, and AUC have been computed and the results point out a high diagnostic potential of the proposed system. This work greatly contributes to addressing the problem of enhancing healthcare delivery especially to rural communities through providing an effective screening tool for early anemia that is inexpensive, accurate and easy to use.