Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: Ph.D. Dissertation
- Language: Farsi
- Document No: 58080 (05)
- University: Sharif University of Technology
- Department: Electrical Engineering
- Advisor(s): Marvasti, Farokh
- Abstract:
- Impulsive noise is a common phenomenon in digital imaging especially in the process of data acquisition and transmission. Hence, impulsive noise removal is considered as a crucial step for any further processing. Utilizing the sparse representation of signals, which has gained attention in recent decades, we first propose a new method called IDT to reconstruct a signal corrupted by noise where both signal and noise are sparse but in different domains. Since signal and noise are assumed to be sparse, impulsive noise removal from images is an important application of our method. Additionally, we propose an iterative method, SIM, to reconstruct simultaneous sparse signals which has various applications in radar and sensor networks. Afterwards, by incorporating this algorithm we modify IDT to remove block sparse noise from sparse signals. Moreover, in recent years, deep neural networks have achieved remarkable performance in different applications of image processing and computer vision such as image enhancement, segmentation, and super-resolution. Finally, in this thesis, we present a deep-learning-based approach to remove impulsive noise from images.
- Keywords:
- Impulsive Noise Removal ; Image Reconstruction ; Sparse Representation ; Iterative Methods ; Deep Learning ; Block Sparse Signal ; Regeneration Signal
Digital Object List
محتواي کتاب
Bookmark