Optimized Regressor Forest for Image Super-Resolution
Chia-Yang Chang, Wei-Chih Tu and Shao-Yi Chien
Abstract
The goal of image super-resolution is to recover missing high frequency details of an image given single or multiple low-resolution images._It is a well-known ill-posed problem and requires mature prior knowledges or enough examples to restore high-quality high-resolution images._Recently, many methods formulate image super-resolution as a regression problem. Input image patches are classified into pre-trained clusters, and cluster-dependent mapping functions are employed to super-resolve input patches._In this paper, for further improving the reconstructed image quality,_an optimized regressor forest framework is proposed, which leverages the discriminative power of random forest._There are three major contributions of the proposed framework._(i) The proposed scheme overturns existing approaches by training the regressors first and learning the way to find the best regressor to avoid quality degradation introduced from the classification outliers._(ii) We propose to employ EM-algorithm to optimize regressors by jointly optimizing the clustering results as well as the regression functions._(iii) In order to find the most appropriate regressor for an input patch at the testing stage,_random forest is adopted to accurately classify patches into their best clusters (regressors)._The experimental results demonstrate that the proposed method generates high-quality high-resolution images and yields state-of-the-art results.
Session
Posters 2
Files
Extended Abstract (PDF, 6M)
Paper (PDF, 13M)
DOI
10.5244/C.30.85
https://dx.doi.org/10.5244/C.30.85
Citation
Chia-Yang Chang, Wei-Chih Tu and Shao-Yi Chien. Optimized Regressor Forest for Image Super-Resolution. In Richard C. Wilson, Edwin R. Hancock and William A. P. Smith, editors, Proceedings of the British Machine Vision Conference (BMVC), pages 85.1-85.12. BMVA Press, September 2016.
Bibtex
@inproceedings{BMVC2016_85,
title={Optimized Regressor Forest for Image Super-Resolution},
author={Chia-Yang Chang, Wei-Chih Tu and Shao-Yi Chien},
year={2016},
month={September},
pages={85.1-85.12},
articleno={85},
numpages={12},
booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
publisher={BMVA Press},
editor={Richard C. Wilson, Edwin R. Hancock and William A. P. Smith},
doi={10.5244/C.30.85},
isbn={1-901725-59-6},
url={https://dx.doi.org/10.5244/C.30.85}
}