Neuromorphic decoding of sample image representations by the boundary-consistent interpolation method

Cover Page

Cite item

Full Text

Open Access Open Access
Restricted Access Access granted
Restricted Access Subscription or Fee Access

Abstract

The paper discusses methods for encoding and decoding large amounts of data using a neuromorphic model based on known neuromechanisms for the perception of visual information. Known mechanisms of the visual system, such as aggregation of counts by receptive fields, central-lateral inhibition, etc., have been studied. A decoding model has been developed that implements the function of simple cells of the primary visual cortex responsible for spatial perception of stimulus contrasts. The proposed decoding model makes it possible to restore local boundaries of objects in an image, while improving the visual quality of images in comparison with the quality of restoration with classical bilinear interpolation.

Full Text

Restricted Access

About the authors

V. A. Kershner

Kotel’nikov Institute of Radio Engineering and Electronics, Russian Academy of Sciences

Author for correspondence.
Email: vladkershner@mail.ru
Russian Federation, Mokhovaya Str., 11, Build. 7, Moscow, 125009

References

  1. Lu Z., Huang D., Bai L. et al. // arXiv preprint arXiv:2304.13023. 2023. https://doi.org/10.48550/arXiv.2304.13023
  2. Pinkston J. T. // IEEE Trans. 1969. V. IT-15. № 1 P. 66. https://doi.org/10.1109/TIT.1969.1054274
  3. Milner D., Goodale M. The Visual Brain in Action. Oxford: Univ. Press, 2006. https://doi.org/10.1093/acprof: oso/9780198524724.001.0001
  4. Antsiperov V., Kershner V. // Pattern Recognition Applications and Methods, ICPRAM 2021–2022. Lecture Notes in Computer Sci. P. 13822. Cham: Springer, 2023. https://doi.org/10.1007/978-3-031-24538-1_3
  5. Yang M., Sun X., Jia F. et al. // Polymers. 2022. V. 14. № 10. Р. 2019. https://doi.org/10.3390/polym14102019
  6. Keeler H. P. Notes on the Poisson Point Process. Technical Report. Berlin: Weierstrass Inst. 2016. 36 p. https://hpaulkeeler.com/wp-content/uploads/2018/08/PoissonPointProcess.pdf
  7. Antsiperov V. // Proc. 11th Int. Conf. on Pattern Recognition Applications and Methods – ICPRAM. 3–5 Feb. 2022. Setúbal: SciTePress – Science and Technology Publ., 2022. P. 354. https://doi.org/10.5220/0010836800003122
  8. Latecki L. J., Lakamper R., Eckhardt T. // Proc. IEEE Conf. Computer Vision and Pattern Recognition, CVPR-2000. Hilton Head Island. 15 Jun. N.Y.: IEEE, 2000. P. 424. https://doi.org/10.1109/CVPR.2000.855850
  9. Hubel D. H., Wiesel T. N. Brain and Visual Perception: The Story of a 25-year Collaboration. Oxford: Univ. Press, 2004. https://doi.org/10.1016/0001-6918(64)90136-2
  10. Keller A. J., Roth M. M., Scanziani M. // Nature. 2020. V. 582. № 7813. Р. 545. https://doi.org/10.1038/s41586-020-2319-4
  11. Hoon M., Okawa H., Santina L. D., Wong R. O. // Progress in Retinal and Eye Research. 2014. V. 42. Р. 44. https://doi.org/10.1016/j.preteyeres.2014.06.003
  12. Antsiperov V. // Proc. 12th Int. Conf. on Pattern Recognition Applications and Methods (ICPRAM 2023). Lisbon. 22–24 Feb. 2023. Setúbal: SciTePress – Science and Technology Publ., 2023. P. 517. https://doi.org/10.5220/0011792800003411
  13. Fish J., Wagner G. J., Keten S. // Nature Mater. 2021. V. 20. № 6. Р. 774. https://doi.org/10.1038/s41563-020-00913-0
  14. Ranstam J., Cook J. A. // J. British Surgery. 2018. V. 105. № 10. Р. 1348. https://doi.org/10.1002/bjs.10895
  15. Tam W. S., Kok C. W., Siu W. C. // J. Electron. Imaging. 2010. V. 19. № 1. Р. 013011. https://doi.org/10.1117/1.3358372
  16. Marr D., Hildreth E. // Proc. Royal Society of London. Ser. B. Biol Sci. 1980. V. 207. № 1167. P. 187. https://doi.org/10.1098/rspb.1980.0020. PMID: 6102765.
  17. Yu S., Zhang R., Wu Sh. et al. // Biomedical Engineering Online. 2013. V. 12. Р. 1. https://doi.org/10.1186/1475-925X-12-102

Supplementary files

Supplementary Files
Action
1. JATS XML
Download
2. Fig. 1. Representation of an image based on a sample of samples (sample representation): a – original image “butterfly-19” [8], b – sample representation of 10 million samples.

Download (67KB)
Indexing metadata
3. Fig. 2. Partitioning of the image surface Ω by a system of receptive ON-fields with square carriers ∆k ∪ ∆s located at the nodes of a regular square lattice.

Download (107KB)
Indexing metadata
4. Fig. 3. Illustration of the coding procedure (11) on a 50×50 receptive field grid of a selective representation of the image “butterfly-19” [8] (see Fig. 1): a – selective representation, b – RP with non-zero values ​​of δ, ON responses (δ > 0) are highlighted in white, OFF responses (δ < 0) are highlighted in black.

Download (99KB)
Indexing metadata
5. Fig. 4. The reconstructed (decoded) image “butterfly-19” [8] (see Fig. 1), defined by the selective representation {} on the 50×50 lattice: a – smoothed image, decoded only using the “smooth” part {nk} of the code, b – interpolation along the edges defined by the details {δk}.

Download (48KB)
Indexing metadata

Copyright (c) 2024 Russian Academy of Sciences