Ananna, ChaimaAnanna, IsmahaneZemmar, Amar2023-11-272023-11-272023http://dspace.univ-oeb.dz:4000/handle/123456789/17057Neurodegenerative diseases are among the most common diseases of our time, Parkinson is our primary goal in this thesis by identifying it in handwriting using deep learning and convolutional neural networks (CNN), where we used transfer learning (vgg16 vgg19) in addition to two other types of convolutional neural networks (CNN3, CNN4) With a data set (handPD, NewhandPD, drawing) for Parkinson's disease by handwriting. The outcomes demonstrated that for (VGG16 and VGG19, our accuracy was excellent and very reasonable, reaching 100% in Drawing dataset. The accuracy of the CNN4 and CNN3 models was 90, 74% in handPD dataset and 96, 29 % in NewhandPD, respectively. the results showed that we can analyze the handwriting of Parkinson's disease using these models.enOther