Presentation | 2014-07-09 high speed super-resolution based on compressed sensing Tomohiro YAMASHITA, Jaehoon YU, Yoshinori TAKEUCHI, Masaharu IMAI, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | This paper proposes a speed-up superresolution(SR) method based on compressed sensing, where SR is a class of techniques to generate a high resolution image from a single or multiple low resolution images. Because it is an ill-posed problem to generate more information than the information of original data, conventional methods for SR adopt machine learning or database based schemes to estimate a plausible high resolution image. However, these types of approaches tend to need a large size of memory and a complicated estimation process. The SR method based on compressed sensing, used in this paper, regards a low resolution image as the result of a compressed form of sensing, and solves the SR problem as a optimization problem of data reconstruction. To realize a speed-up of the optimization process, we use the quasi-Newton method instead of the ROMP algorithm used in the previous research. In the experimental result using standard image database, the proposed method achieved 1 to 4 dB higher PSNR than conventional image magnification methods and showed almost equivalent magnification performance with the method proposed in the previous research, and realized 2 to 21 times speed-up of processing performance. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | superresolution / compressed sensing / quasi-Newton method |
Paper # | CAS2014-16,VLD2014-25,SIP2014-37,MSS2014-16,SIS2014-16 |
Date of Issue |
Conference Information | |
Committee | SIS |
---|---|
Conference Date | 2014/7/2(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 | Smart Info-Media Systems (SIS) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | high speed super-resolution based on compressed sensing |
Sub Title (in English) | |
Keyword(1) | superresolution |
Keyword(2) | compressed sensing |
Keyword(3) | quasi-Newton method |
1st Author's Name | Tomohiro YAMASHITA |
1st Author's Affiliation | Department of Information Systems Engineering, Graduate School of Information Science and Technology, Osaka University() |
2nd Author's Name | Jaehoon YU |
2nd Author's Affiliation | Department of Information Systems Engineering, Graduate School of Information Science and Technology, Osaka University |
3rd Author's Name | Yoshinori TAKEUCHI |
3rd Author's Affiliation | Department of Information Systems Engineering, Graduate School of Information Science and Technology, Osaka University |
4th Author's Name | Masaharu IMAI |
4th Author's Affiliation | Department of Information Systems Engineering, Graduate School of Information Science and Technology, Osaka University |
Date | 2014-07-09 |
Paper # | CAS2014-16,VLD2014-25,SIP2014-37,MSS2014-16,SIS2014-16 |
Volume (vol) | vol.114 |
Number (no) | 126 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |