Echo State Network for Features Extraction and Segmentation of Tomography Images

Petia Koprinkova-Hristova1, Ivan Georgiev1, 2 and Miryana Raykovska1

  1. Institute of Information and Communication technologies, Bulgarian Academy of Sciences
    Acad. G. Bonchev str. bl.2, Sofia 1113, Bulgaria
    petia.koprinkova@iict.bas.bg, ivan.georgiev@parallel.bas.bg, miriana.raykovska@iict.bas.bg
  2. Institute of Mathematics and Informatics, Bulgarian Academy of Sciences
    Acad. G. Bonchev str. bl.8, Sofia 1113, Bulgaria

Abstract

The paper proposes a novel approach for gray scale images segmentation. It is based on multiple features extraction from a single feature per image pixel, namely its intensity value, via a recurrent neural network from the reservoir computing family - Echo state network. The preliminary tests on the benchmark gray scale image Lena demonstrated that the newly extracted features - reservoir equilibrium states - reveal hidden image characteristics. In present work the developed approach was applied to a real life task for segmentation of a 3D tomography image of a of bone whose aim was to explore the object’s internal structure. The achieved results demonstrated the novel approach allows for clearer revealing the details of the bone internal structure thus supporting further tomography image analyses.

Key words

reservoir computing, Echo state network, intrinsic plasticity, gray image, segmentation, tomography

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS230128045K

Publication information

Volume 21, Issue 1 (January 2024)
Year of Publication: 2024
ISSN: 2406-1018 (Online)
Publisher: ComSIS Consortium

Full text

DownloadAvailable in PDF
Portable Document Format

How to cite

Koprinkova-Hristova, P., Georgiev, I., Raykovska, M.: Echo State Network for Features Extraction and Segmentation of Tomography Images. Computer Science and Information Systems, Vol. 21, No. 1, 379–393. (2024), https://doi.org/10.2298/CSIS230128045K