The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Na super-resolução baseada em reconstrução, uma imagem de alta resolução é estimada usando múltiplas imagens de baixa resolução com desalinhamentos de subpixels. Portanto, quando apenas uma imagem de baixa resolução está disponível, geralmente é difícil obter uma imagem favorável. Esta carta propõe um método para superar esta dificuldade para super-resolução de imagem única. Em nosso método, após interpolar valores de pixel em locais de subpixel patch por patch por regressão de vetor de suporte, na qual amostras de aprendizagem são coletadas dentro de uma determinada imagem com base em semelhanças locais, resolvemos o problema de reconstrução regularizada com um número suficiente número de restrições. Experimentos de avaliação foram realizados para imagens artificiais e naturais, e as imagens de alta resolução obtidas indicam favoravelmente os componentes de alta frequência, juntamente com PSNRs melhorados.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Atsushi YAGUCHI, Tadaaki HOSAKA, Takayuki HAMAMOTO, "Image Quality Enhancement for Single-Image Super Resolution Based on Local Similarities and Support Vector Regression" in IEICE TRANSACTIONS on Fundamentals,
vol. E94-A, no. 2, pp. 552-554, February 2011, doi: 10.1587/transfun.E94.A.552.
Abstract: In reconstruction-based super resolution, a high-resolution image is estimated using multiple low-resolution images with sub-pixel misalignments. Therefore, when only one low-resolution image is available, it is generally difficult to obtain a favorable image. This letter proposes a method for overcoming this difficulty for single- image super resolution. In our method, after interpolating pixel values at sub-pixel locations on a patch-by-patch basis by support vector regression, in which learning samples are collected within the given image based on local similarities, we solve the regularized reconstruction problem with a sufficient number of constraints. Evaluation experiments were performed for artificial and natural images, and the obtained high-resolution images indicate the high-frequency components favorably along with improved PSNRs.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.E94.A.552/_p
Copiar
@ARTICLE{e94-a_2_552,
author={Atsushi YAGUCHI, Tadaaki HOSAKA, Takayuki HAMAMOTO, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Image Quality Enhancement for Single-Image Super Resolution Based on Local Similarities and Support Vector Regression},
year={2011},
volume={E94-A},
number={2},
pages={552-554},
abstract={In reconstruction-based super resolution, a high-resolution image is estimated using multiple low-resolution images with sub-pixel misalignments. Therefore, when only one low-resolution image is available, it is generally difficult to obtain a favorable image. This letter proposes a method for overcoming this difficulty for single- image super resolution. In our method, after interpolating pixel values at sub-pixel locations on a patch-by-patch basis by support vector regression, in which learning samples are collected within the given image based on local similarities, we solve the regularized reconstruction problem with a sufficient number of constraints. Evaluation experiments were performed for artificial and natural images, and the obtained high-resolution images indicate the high-frequency components favorably along with improved PSNRs.},
keywords={},
doi={10.1587/transfun.E94.A.552},
ISSN={1745-1337},
month={February},}
Copiar
TY - JOUR
TI - Image Quality Enhancement for Single-Image Super Resolution Based on Local Similarities and Support Vector Regression
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 552
EP - 554
AU - Atsushi YAGUCHI
AU - Tadaaki HOSAKA
AU - Takayuki HAMAMOTO
PY - 2011
DO - 10.1587/transfun.E94.A.552
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E94-A
IS - 2
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - February 2011
AB - In reconstruction-based super resolution, a high-resolution image is estimated using multiple low-resolution images with sub-pixel misalignments. Therefore, when only one low-resolution image is available, it is generally difficult to obtain a favorable image. This letter proposes a method for overcoming this difficulty for single- image super resolution. In our method, after interpolating pixel values at sub-pixel locations on a patch-by-patch basis by support vector regression, in which learning samples are collected within the given image based on local similarities, we solve the regularized reconstruction problem with a sufficient number of constraints. Evaluation experiments were performed for artificial and natural images, and the obtained high-resolution images indicate the high-frequency components favorably along with improved PSNRs.
ER -