Loading...
Search
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
Digital Object List
Bookmark
Review List
814 viewed
کاربردهای نمایش های تنک در افزایش تفکیک پذیری تصاویر
صحرایی اردکان، مجتبی Sahraee-Ardakani, Mojtaba
Sparse Representation and its Application in Image Super-resolution
Sahraee-Ardakani, Mojtaba | 2013
662
Viewed
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 44538 (05)
- University: Sharif University of Technology
- Department: Electrical Engineering
- Advisor(s): Babaie-Zadeh, Massoud
- Abstract:
- Sparse signal representations and its applications has been a hot topic of research in recent years. It has been demonstrated that sparsity prior can be effectively used as a regularization term to solve many of the inverse problems. One of these problems in which sparse representations have been used is image super-resolution (SR). SR is the problem of finding a high resolution (HR) image from one or several low resolution (LR) images. In this dissertation, we have focused on the problem of finding a HR image from only one LR image which is known as example-based SR. There are two kinds of methods for example-based SR: the methods which use neighborhood embedding and the methods which use sparse representations.The methods which use sparse representations are newer, require less memory and they are much faster. In order to solve the example-based SR problem, we need to have some information other than those of the LR image itself. This information is obtained from pairs of HR and LR image patches sampled from some example images we already have. In methods which use neighborhood embedding, these pairs of image patches are directly used. But in methods based on sparse representations, we use these image patches to train two dictionaries which are used to represents HR and LR image patches. In this dissertation, an example-based SR algorithm proposed by Yang is discussed in detail in chapter 2. Subsequently one of the problems of Yang’s algorithm is studied and a general idea to attack this problem is proposed. Based on this general idea we have developed several algorithms for example-based SR which slightly differ from each other. It is shown that dictionaries trained by some of these algorithms work better than dictionaries trained by Yang’ algorithm, in the sense that the error of HR images produced by some of our methods are a little less than error of HR images produced by Yang’s algorithm using PSNR as criterion
- Keywords:
- Sparse Representation ; Super-Resolution ; Dictionary Learning ; Neighbor Embedding
-
محتواي کتاب
- view