DEEP LEARNING APPROACH USING CNN MODELS TOWARDS EFFECTUAL TREATMENT OF PARKINSON DISEASE /
WISDOM CHIGOZIE AGUNTA SUPERVISOR ASSOC. PROF. DR. HÜSEYIN ÖZTOPRAK
- xiii, 65 sheets; 31 cm. Includes CD
Thesis (MSc) - Cyprus International University. Institute of Graduate Studies and Research Information Technologies Department
Includes bibliography (sheets 60-65)
ABSTRACT This thesis proposal aims to explore the potential of utilizing convolutional neural network, in particular, deep learning techniques (CNN) models, to improve the efficacy of treatment options for individuals diagnosed with Parkinson's Disease (PD). The proposed research will focus on developing and evaluating advanced CNN models that can accurately predict the progression of PD and assist in the selection of personalized treatment options. A comprehensive dataset of imaging and clinical data from individuals with PD will be collected and used to train and evaluate the performance of the proposed CNN models. The proposed research will also investigate the incorporation of state-of-the-art CNN architectures such as MobileNet_v2 to enhance the capabilities of the models in accurately predicting PD progression. Accuracy, precision, recall, and F1 score will be used to objectively assess the models' performance. The results will then be studied to ascertain whether CNN models have the potential to enhance PD treatment options. The results of this investigation will have important ramifications for the development of more accurate and personalized treatment options for PD and will contribute to the ongoing efforts to improve the management and treatment of this debilitating disorder. Keywords: Clinical data, CNN models, Deep learning, Development, Effectual, Imaging data, Implications, Management, MobileNet, MobileNet_v2, Ongoing efforts, Personalized medicine, Personalized treatment, Parkinson's disease, Progression prediction, Treatment.