| Presentation | 2013-03-14 Self-learning Super Resolution by l_2-norm Daiki AZUMA, Kazu MISHIBA, Masaaki IKEHARA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In this paper, we propose a single image super resolution technique by l_2 approximation without any training. Recently it has been proposed that image super resolution by using sparse representation. However it has to make a dictionary from tremendous low and high resolution image pairs to apply in any kinds of inputs. Moreover this algorithm takes a long time to train and requires the dictionaries to learn in advance. In our method, the high- and low- resolution dictionaries are made from random selected patch pairs of input and its reduced image, respectively. Then the input image is approximated by low-resolution dictionary in the sense of l_2 and its coefficients are used to reconstruct the high-resolution image. Therefore the proposed method does not need any training to make dictionaries and any time consuming computation for approximations. In some example, our method has comparable quality to the conventional method. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | super resolution / sparse coding / l_2-norm minimization / self-similarity |
| Paper # | CAS2012-106,SIP2012-137,CS2012-112 |
| Date of Issue |
| Conference Information | |
| Committee | CAS |
|---|---|
| Conference Date | 2013/3/7(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Circuits and Systems (CAS) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Self-learning Super Resolution by l_2-norm |
| Sub Title (in English) | |
| Keyword(1) | super resolution |
| Keyword(2) | sparse coding |
| Keyword(3) | l_2-norm minimization |
| Keyword(4) | self-similarity |
| 1st Author's Name | Daiki AZUMA |
| 1st Author's Affiliation | Department of Electronics and Electrical Engineering, Keio University() |
| 2nd Author's Name | Kazu MISHIBA |
| 2nd Author's Affiliation | Department of Electrical and Electronic Engineering, Tottori University |
| 3rd Author's Name | Masaaki IKEHARA |
| 3rd Author's Affiliation | Department of Electronics and Electrical Engineering, Keio University |
| Date | 2013-03-14 |
| Paper # | CAS2012-106,SIP2012-137,CS2012-112 |
| Volume (vol) | vol.112 |
| Number (no) | 484 |
| Page | pp.pp.- |
| #Pages | 5 |
| Date of Issue | |