15866

14166

كارشناسي ارشد

مخابرات سيستم

اصفهان : دانشگاه صنعتي اصفهان

1399

چهارده،97ص، مصور، جدول، نمودار

محمدعلي خسروي فرد

محمدرضا احمدزاده، بهزاد نظري

1399/07/08

كتابنامه

مهندسي برق

مهندسي برق و كامپيوتر

1399/07/27

2632455

Application of Convolutional Sparse Coding in Deep Image Representation Zahra Sadat Ghoreyshi zs ghoreyshi@ec iut ac ir Jul 2020 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language Persian Supervisor Dr Mohammadali Khosravifard Abstract While Convolutional Neural Networks CNN have proven their competence in various signal and image processingtasks the theory of CNN in the forward pass is still lacking In parallel recent advances in sparse coding have attracted muchattention to Convolutional Sparse Coding CSC Providing a global sparse model CSC can overcome several limitationsof the patch based sparse model ML CSC is emerged from the cascade of CSC layers demonstrating the close connectionbetween CNN forward pass and sparse coding This connection brings a fresh view to CNN under simple local sparsityconditions This study proposes a new structure of ML CSC network along with a new adaptive approach to design a globalstructural dictionary This approach uses a multi dictionary learning model and the dictionary optimization algorithm learnsin wavelet sub bands This algorithm improves the adaptation and complexity of atoms in the trained set of dictionariescompared to the single dictionary structure Furthermore because the pooling operation loses some location information we propose a structure for deep network without using the pooling layer and demonstrate the advantage of the proposedalgorithm for image denoising in terms of performance and convergence Key Words Sparse Coding Dictionary learning Wavelet Deep Convolutional Neural Network

محمدعلي خسروي فرد

محمدرضا احمدزاده، بهزاد نظري