Presentation 2015-03-06
High Speed Sparse-coding Superresolution by Multiple Degenerated Basis Dictionaries
Naoki MOTOHASHI, Satoshi NAKAMURA, Toshihiro SUZUKI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Single-frame Superresolution (Single-frame SR) is a technique for improving the quality of low resolution images, which restores lost fine details (e.g., textures) by adding high frequency components extracted from other images. In the field of Single-frame SR, Example-based Superresolution (Example-based SR) has been actively studied for many years because of its high restoration capability. Recently Example-based SR using sparse coding (ScSR) gets increasingly popular because it can restore the lost fine details accurately while the processing time is significantly reduced compared to the ordinary Example-based SR. However, due to its high computational complexity, it is still required to accelerate the algorithm. To this end, we propose a new technique to accelerate ScSR algorithm by splitting the dictionary used in ScSR to multiple sub-dictionaries to reduce the number of matchings between input patches, i.e., small blocks extracted from an input low resolution image, and bases in the dictionary. Experimental results show that the proposed technique can reduce the processing time by more than 50% without compensating the restoration accuracy.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) example-based Superresolution / sparse-coding / basis dictionaries
Paper # SIS2014-106
Date of Issue

Conference Information
Committee SIS
Conference Date 2015/2/26(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 Sparse-coding Superresolution by Multiple Degenerated Basis Dictionaries
Sub Title (in English)
Keyword(1) example-based Superresolution
Keyword(2) sparse-coding
Keyword(3) basis dictionaries
1st Author's Name Naoki MOTOHASHI
1st Author's Affiliation Ricoh Institute of Technology Ricoh Company, LTD.()
2nd Author's Name Satoshi NAKAMURA
2nd Author's Affiliation Ricoh Institute of Technology Ricoh Company, LTD.
3rd Author's Name Toshihiro SUZUKI
3rd Author's Affiliation Ricoh Institute of Technology Ricoh Company, LTD.
Date 2015-03-06
Paper # SIS2014-106
Volume (vol) vol.114
Number (no) 496
Page pp.pp.-
#Pages 6
Date of Issue