Implementation of U-net architecture in image segmentation of cervical cells

dc.contributor.advisorKovács, László
dc.contributor.authorYagublu, Kanan
dc.contributor.departmentDE--Informatikai Karhu_HU
dc.description.abstractAim of this project was to develop computer-based, precisely saying deep-learning-based diagnoses of the Pap smear test. The result of this study is expected to give us segmented image samples, after processing the raw input samples and provide us with a practical system that will be helpful for doctors for better diagnoses. The whole study is consisted of mainly two parts preparing data and training the fully convolutional neural network, within those parts we had several stages. As a consequence of this research, it can be concluded that modern world technological tools like AI, can be really useful in terms of medical assistance, helping to detect and diagnose cervical precancer cells before its lethal level of disease. For this purpose in this paper U-net deep learning architecture applied for detecting precancerous cells in the human cervix or colon. Regarding architecture, U-net fully described and explained in the paper and regarding the data, how data obtained, from where and how preprocessed for feeding neural network discussed. You will also see the result of the training process as the model will segment the input image.hu_HU
dc.description.courseComputer Science Engineeringhu_HU
dc.subjectDeep learninghu_HU
dc.subjectImage segmentationhu_HU
dc.subjectCell detectionhu_HU
dc.subject.dspaceDEENK Témalista::Informatikahu_HU
dc.titleImplementation of U-net architecture in image segmentation of cervical cellshu_HU