Learning image super resolution from joint examples

Image super-resolution (SR) aims to estimate of a high-resolution (HR) image from low-resolution (LR) input. Image priors are commonly learned to regularize the ill-posed SR problem, either using external LR-HR pairs or internal similar patterns repeating across di erent scales. We propose joint SR to adaptively combine the advantages of both external and internal SR. We de ne the two loss functions using sparse coding and epitomic matching, respectively. A corresponding adaptive weight is constructed to balance their e ect according to the reconstruction errors. Various image results demonstrate the e ectiveness of the proposed method over the existing state-of-the-art methods, which is also veri ed by our subject evaluation experiment.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Wang, Zhangyang

Contributors dc:contributor

• Huang, Thomas S.

• super-resolution
• example-based learning
• sparse coding
• epitomic matching
• subject evaluation

Statement dc:rights

• Copyright 2014 Zhangyang Wang

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/73058

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/73058