A MobileNetV2-Based Hybrid Deep Learning Framework for HighAccuracy Monkeypox Detection and Real-Time Web Application

Abstract

The improved deep learning-based system for automated monkeypox detection utilizing transfer learning techniques is presented in this research. The categorization of skin lesion photos is assessed using many pre-trained convolutional neural network models, such as VGG16, ResNet50, VGG19, and MobileNetV2. A Hybrid MobileNetV2 model is suggested to enhance diagnostic performance. It achieves better accuracy than current models by combining enhanced feature extraction and classification capabilities. Additionally, a Flask-based web application connected with user authentication is added to the system to enable secure and real-time diagnostics. Experiments show that the suggested hybrid model performs exceptionally well, which makes it appropriate for quick, affordable, and easily accessible monkeypox detection, particularly in settings with limited resources.