ANALYSIS AND DETECTION OF MONKEYPOX USING MACHINE LEARNING

Abstract

Monkeypox is an infectious disease with very important implications for public health. Therefore, it is not just the development of methods for detection that could be quickly carried out; they also must have enough accuracy to manage and control it effectively. This thesis carries out an elaborate kind of study on the use of deep learning techniques in the development and evaluation of computer software purposed for the detection of monkeypox. This will be done by developing and testing deep learning architectures for identifying monkeypox using skin lesion images, which can enable the quick and accurate detection of any case of monkeypox. We develop different deep learning architectures and strategies to optimize image processing techniques that extract meaningful data out of skin lesion images in our study. Key performance metrics, like accuracy, sensitivity, and specificity, are given much emphasis in the assessment of the effectiveness of the models. This shows a very good recall for our deep learning models: EfficientNetB3 89.8%, VGG16 90.4%, and ResNet50 93.8%. By using the Monkeypox Skin Lesion Dataset 2.0 (MSLD 2.0) only, we remove all questions about reliability or validity in our results. This means that robust training and evaluation of the model can be performed. Our study helps to improve monkeypox detection systems, automate such processes for public health outcomes, and ensure the intervention accorded is timely in communities affected by the disease. Those could be some great insights on the performance of deep learning models for monkeypox detection and how they underline matter-of-factly the strengths and weaknesses of different architectures. We clarify more on the role of deep learning by further rigorous experimentation and continuous analysis in aiming to enhance the accuracy and efficiency of the monkeypox detection system. We foster advancements in deep learning methodologies for infectious disease detection through focusing on collaborative research efforts, sharing knowledge across the scientific community, and translating into better health care at a global level.