Image interpolation using cosine-sine modulated filter bank

Shunsuke Iwamura, Seisuke Kyochi, Masaaki Ikehara, Naoto Kaneko

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper proposes a new image interpolation technique by using direction-aware transform. In literature, it is known that Cosine-Sine modulated filter banks (CSMFB) can be regarded as alternative dual-tree complex wavelet transform, therefore, CSMFB has rich directional selectivity. Our interpolation procedure consists of two stages. First, we generate low frequency image and six directional images from the input image by using 4-channel CSMFB. Then each image is interpolated along with each direction. The simulation results show that our method can produce high visual quality images and improves some artifacts. Moreover, our method is quite feasible from the view point of hardware implementation since there is no branching procedure.

Original languageEnglish
Title of host publicationProceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages13-16
Number of pages4
ISBN (Print)9781509020768
DOIs
Publication statusPublished - 2016 Feb 17
Event1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010 - Foz do Iguacu, Brazil
Duration: 2010 Feb 242010 Feb 26

Other

Other1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010
CountryBrazil
CityFoz do Iguacu
Period10/2/2410/2/26

Fingerprint

Filter banks
Interpolation
Wavelet transforms
Image quality
Hardware

Keywords

  • cosine-sine modulated filter banks
  • Direction-aware transform
  • image interpolation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Iwamura, S., Kyochi, S., Ikehara, M., & Kaneko, N. (2016). Image interpolation using cosine-sine modulated filter bank. In Proceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010 (pp. 13-16). [7410128] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/LASCAS.2010.7410128

Image interpolation using cosine-sine modulated filter bank. / Iwamura, Shunsuke; Kyochi, Seisuke; Ikehara, Masaaki; Kaneko, Naoto.

Proceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010. Institute of Electrical and Electronics Engineers Inc., 2016. p. 13-16 7410128.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Iwamura, S, Kyochi, S, Ikehara, M & Kaneko, N 2016, Image interpolation using cosine-sine modulated filter bank. in Proceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010., 7410128, Institute of Electrical and Electronics Engineers Inc., pp. 13-16, 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010, Foz do Iguacu, Brazil, 10/2/24. https://doi.org/10.1109/LASCAS.2010.7410128
Iwamura S, Kyochi S, Ikehara M, Kaneko N. Image interpolation using cosine-sine modulated filter bank. In Proceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010. Institute of Electrical and Electronics Engineers Inc. 2016. p. 13-16. 7410128 https://doi.org/10.1109/LASCAS.2010.7410128
Iwamura, Shunsuke ; Kyochi, Seisuke ; Ikehara, Masaaki ; Kaneko, Naoto. / Image interpolation using cosine-sine modulated filter bank. Proceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010. Institute of Electrical and Electronics Engineers Inc., 2016. pp. 13-16
@inproceedings{747e52bf48ed4705872af1fdf131226a,
title = "Image interpolation using cosine-sine modulated filter bank",
abstract = "This paper proposes a new image interpolation technique by using direction-aware transform. In literature, it is known that Cosine-Sine modulated filter banks (CSMFB) can be regarded as alternative dual-tree complex wavelet transform, therefore, CSMFB has rich directional selectivity. Our interpolation procedure consists of two stages. First, we generate low frequency image and six directional images from the input image by using 4-channel CSMFB. Then each image is interpolated along with each direction. The simulation results show that our method can produce high visual quality images and improves some artifacts. Moreover, our method is quite feasible from the view point of hardware implementation since there is no branching procedure.",
keywords = "cosine-sine modulated filter banks, Direction-aware transform, image interpolation",
author = "Shunsuke Iwamura and Seisuke Kyochi and Masaaki Ikehara and Naoto Kaneko",
year = "2016",
month = "2",
day = "17",
doi = "10.1109/LASCAS.2010.7410128",
language = "English",
isbn = "9781509020768",
pages = "13--16",
booktitle = "Proceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Image interpolation using cosine-sine modulated filter bank

AU - Iwamura, Shunsuke

AU - Kyochi, Seisuke

AU - Ikehara, Masaaki

AU - Kaneko, Naoto

PY - 2016/2/17

Y1 - 2016/2/17

N2 - This paper proposes a new image interpolation technique by using direction-aware transform. In literature, it is known that Cosine-Sine modulated filter banks (CSMFB) can be regarded as alternative dual-tree complex wavelet transform, therefore, CSMFB has rich directional selectivity. Our interpolation procedure consists of two stages. First, we generate low frequency image and six directional images from the input image by using 4-channel CSMFB. Then each image is interpolated along with each direction. The simulation results show that our method can produce high visual quality images and improves some artifacts. Moreover, our method is quite feasible from the view point of hardware implementation since there is no branching procedure.

AB - This paper proposes a new image interpolation technique by using direction-aware transform. In literature, it is known that Cosine-Sine modulated filter banks (CSMFB) can be regarded as alternative dual-tree complex wavelet transform, therefore, CSMFB has rich directional selectivity. Our interpolation procedure consists of two stages. First, we generate low frequency image and six directional images from the input image by using 4-channel CSMFB. Then each image is interpolated along with each direction. The simulation results show that our method can produce high visual quality images and improves some artifacts. Moreover, our method is quite feasible from the view point of hardware implementation since there is no branching procedure.

KW - cosine-sine modulated filter banks

KW - Direction-aware transform

KW - image interpolation

UR - http://www.scopus.com/inward/record.url?scp=84964858030&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84964858030&partnerID=8YFLogxK

U2 - 10.1109/LASCAS.2010.7410128

DO - 10.1109/LASCAS.2010.7410128

M3 - Conference contribution

SN - 9781509020768

SP - 13

EP - 16

BT - Proceedings - 2010 1st IEEE Latin American Symposium on Circuits and Systems, LASCAS 2010

PB - Institute of Electrical and Electronics Engineers Inc.

ER -