Deep Learning Approach for Real-Time Monkeypox Diagnosis

Monkeypox, a re-emerging zoonotic viral disease, is often misdiagnosed due to its clinical resemblance to other skin conditions such as chickenpox and measles. Traditional diagnostic methods like clinical observation and PCR testing, while accurate, are resource-intensive, time- consuming, and inaccessible in under-resourced regions. To address these limitations, this paper proposes a Monkeypox Detection System—a lightweight, web-based, AI-powered diagnostic tool that leverages deep learning for rapid identification. The system integrates a fine-tuned VGG19 convolutional neural network to classify skin lesion images into monkeypox or non-monkeypox categories with high accuracy. Implemented using Flask, the platform supports real-time prediction, voice-command navigation, PDF report generation, and interactive data visualization. User authentication and prediction history enhance reliability, while visual analytics aid medical interpretation. The modular architecture ensures scalability for integration with healthcare systems and expansion to multiple skin diseases. Experimental results demonstrate that the system achieves reliable performance in real-world scenarios, offering a cost-effective and accessible solution for early diagnosis. This work contributes to AI-enabled preventive healthcare, particularly in outbreak-prone and resource-limited settings.