Now showing items 1-2 of 2

  • Semantic segmentation of diabetic retinopathy lesions, using a UNET with pretrained encoder.

    H.M.U., School of Engineering (ScENG) MSc in Informatics & Multimedia
    Authors: Theodoropoulos, Dimitrios
    Thesis advisor: Marias, Konstantinos
    Publication Date: 2022-01-19
    With the emergence of Deep Learning nowadays, a lot of novel architectures have been devised to perform tasks such as classification, detection, segmentation etc. In Medical Imaging and especially in Ophthalmology, the ...
  • Inverse physical parameter estimation of vibrating membranes from sound using deep neural networks.

    H.M.U., School of Music and Optoacoustic Technologies (SMOT), MSc in Sound and Music Technologies
    Authors: Starakis, Michail
    Thesis advisor: Alexandraki, Chrysoula; Zervas, Panagiotis
    Publication Date: 2022-06-02
    Inverse problems in the field of acoustics can be valuable tools for retrieving information through the produced sound. At the same time, the rapid development of deep machine learning techniques and the remarkable ...