Photoacoustic image denoising using dictionary learning

Syahril Siregar, Yoshifumi Saijo

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

Abstract

Photoacoustic (PA) imaging is the biomedical imaging modality to visualize the biological object with high contrast, high spatial and temporal resolutions. The PA image is degraded due to several parameters such as random noise, frequency, transducer, and laser components. A band-pass filter does not completely remove the noise since the noise is distributed in the bandwidth frequency. In this paper, we propose noise removal method for PA image by applying dictionary learning method. The algorithm is applied to PA images of micropipe filled carbon nanotube and in vivo mice ear. We estimated the optimum input parameters to implement dictionary learning denoising method on PA image. Our results declared that the proposed denoising method using dictionary learning enhances the quality of PA image.

Original languageEnglish
Title of host publicationProceedings of 2018 3rd International Conference on Biomedical Signal and Image Processing, ICBIP 2018
PublisherAssociation for Computing Machinery
Pages34-37
Number of pages4
ISBN (Electronic)9781450364362
DOIs
Publication statusPublished - 2018 Aug 22
Event3rd International Conference on Biomedical Signal and Image Processing, ICBIP 2018 - Seoul, Korea, Republic of
Duration: 2018 Aug 222018 Aug 24

Publication series

NameACM International Conference Proceeding Series

Other

Other3rd International Conference on Biomedical Signal and Image Processing, ICBIP 2018
CountryKorea, Republic of
CitySeoul
Period18/8/2218/8/24

Keywords

  • Dictionary learning
  • Image denoising
  • PSNR
  • Photoacoustic image

ASJC Scopus subject areas

  • Software
  • Human-Computer Interaction
  • Computer Vision and Pattern Recognition
  • Computer Networks and Communications

Fingerprint Dive into the research topics of 'Photoacoustic image denoising using dictionary learning'. Together they form a unique fingerprint.

Cite this