Publication detail

Diffracted Image Restoration: A Machine learning approach

KOUDELKA, V. DEL RIO BOCIO, C. RAIDA, Z.

Original Title

Diffracted Image Restoration: A Machine learning approach

English Title

Diffracted Image Restoration: A Machine learning approach

Type

conference paper

Language

en

Original Abstract

Image restoration issues are closely connected with imaging systems, where image resolution is limited by diffraction phenomenon. The presented work is motivated by the super acuity of the Human vision, where the restoration step is implemented by some kind of parallel processor unit - neural network. The de-convolution process is formulated as a machine learning problem and the inverse operator is interpreted as a connectionist model.

English abstract

Image restoration issues are closely connected with imaging systems, where image resolution is limited by diffraction phenomenon. The presented work is motivated by the super acuity of the Human vision, where the restoration step is implemented by some kind of parallel processor unit - neural network. The de-convolution process is formulated as a machine learning problem and the inverse operator is interpreted as a connectionist model.

Keywords

Diffraction, Image restoration, Imaging, Noise, Sensors, Stability analysis, Training

RIV year

2013

Released

09.09.2013

Publisher

COREP

Location

Torino, Italy

ISBN

978-1-4673-5705-0

Book

Proceedings of 2013 International Conference on Electromagnetics in Advanced Applications

Pages from

931

Pages to

934

Pages count

4

BibTex


@inproceedings{BUT102451,
  author="Vlastimil {Koudelka} and Zbyněk {Raida}",
  title="Diffracted Image Restoration: A Machine learning approach",
  annote="Image restoration issues are closely connected with imaging systems, where image resolution is limited by diffraction phenomenon. The presented work is motivated by the super acuity of the Human vision, where the restoration step is implemented by some kind of parallel processor unit - neural network. The de-convolution process is formulated as a machine learning problem and the inverse operator is interpreted as a connectionist model.",
  address="COREP",
  booktitle="Proceedings of 2013 International Conference on Electromagnetics in Advanced Applications",
  chapter="102451",
  doi="10.1109/ICEAA.2013.6632375",
  howpublished="electronic, physical medium",
  institution="COREP",
  year="2013",
  month="september",
  pages="931--934",
  publisher="COREP",
  type="conference paper"
}