![](bg1.png)
Fuzzy-rule-embedded Reduction Image Construction Method
for Image Enlargement with High Magnification
Hakaru Tamukoh
1
, Noriaki Suetake
2
, Hideaki Kawano
3
, Ryosuke Kubota
4
, Byungki Cha
5
and Takashi Aso
5
1
Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology, Kyushu, Japan
2
Graduate School of Science and Engineering, Yamaguchi University, Yamaguchi, Japan
3
Graduate School of Engineering, Kyushu Institute of Technology, Kyushu, Japan
4
Department of Intelligent System Engineering, Ube National College of Technology, Ube, Japan
5
Faculty of Management and Information Sciences, Kyushu Institute of Information Sciences, Kyushu, Japan
Keywords:
Image Enlargement, Image Reduction, Data Embedding, Fuzzy Inference.
Abstract:
This paper proposes a fuzzy-rule-embedded reduction image construction method for image enlargement. A
fuzzy rule is generated by considering distribution of pixel value around a target pixel. The generated rule is
embedded into the target pixel in a reduction image. The embedded fuzzy rule is used in a fuzzy inference
to generate a highly magnified image from the reduction image. Experimental results, which scale factors
are three and four, show that the proposed method realizes high-quality image enlargement in terms of both
objective and subjective evaluations in comparison with conventional methods.
1 INTRODUCTION
In recent years, high-resolution displays have become
widely used such as high-definition televisions, mo-
bile devices and smart phones. In addition, a 4K
(3840×2160 pixels) resolution already exist in digi-
tal television and digital cinematography, and an 8K
(7680×4320 pixels) resolution will be available as
ultra-high-definition displays in the near future. At
the same time, people can obtain over giga-pixel
images, because high-resolution digital cameras are
widely commoditized. In addition, image- and video-
sharing services become as common all over the
world. To upload image or video to these services,
people have to reduce image size into less than quar-
ter size. Naturally, users require browsing high-
resolution images on the high-resolution displays. To
satisfy this requirement, whole or part of image have
to be enlarged larger than four times in size.
Image reduction and enlargement methods are
very important technologies in sharing and display-
ing images among such devices. Classical image
scaling methods—such as nearest neighbor interpo-
lation (NNI), bilinear interpolation (BLI), and bi-
cubic Interpolation (BCI)—are based on interpola-
tion using kernels (Lin, 1990), (Keys, 1981). These
interpolation-based methods achieve fast smooth im-
age reduction and enlargement; however, once images
are reduced by these methods, they cannot restore the
high-frequency image components lost in the reduc-
tion process, and therefore cannot preserve clearly the
step edges and peaks of an image. This is caused
by the fact that the high-frequencyimage components
beyond the Nyquist frequency cannot be restored us-
ing these simple kernel-based methods. If multiple
images are available, a high-resolution image can be
generated from a set of low-resolution images in the
same scene (Farsiu et al., 2004), but they cannot be
applied to stationary images. To address this prob-
lem, various advanced image enlargement methods
from the single image accompanying the estimation
of the high-frequencycomponent have been proposed
(Greenspan et al., 2000), (Siu and Hung, 2012). The
estimated high-frequency component is overlapped
with a blurred image generated by interpolation based
methods, to generate a high-quality image. However,
estimation of high-frequency component is difficult
when scale factor is over three or four.
In this paper, we propose a fuzzy-rule-embedded
image construction method to generate a reduction-
image in the image reduction process. The embedded
fuzzy rules are used in the proposed fuzzy inference to
generate an enlarged image with high magnification
in the image enlargement process. To show the effec-
tiveness of the proposed method, we compare results
of the proposed method with the conventional meth-
228
Tamukoh H., Suetake N., Kawano H., Kubota R., Cha B. and Aso T..
Fuzzy-rule-embedded Reduction Image Construction Method for Image Enlargement with High Magnification.
DOI: 10.5220/0004851802280233
In Proceedings of the 9th International Conference on Computer Vision Theory and Applications (VISAPP-2014), pages 228-233
ISBN: 978-989-758-003-1
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)