intechnguВестник НГУ. Серия: Информационные технологииVestnik NSU. Series: Information Technologies1818-79002410-0420НГУ10.25205/1818-7900-2021-19-3-50-60intechngu-173Research ArticleСтатьиНовая модель нейрона конца линий, применимая в свёрточных нейронных сетяхNew End-Stopped Complex Cell Model Applicable in Convolutional Neural Networkshttps://orcid.org/0000-0002-6676-0518КугаевскихА. В.KugaevskikhA. V.

 Александр Владимирович Кугаевских, кандидат технических наук 

Новосибирск

 Alexander V. Kugaevskikh, Candidate of Sciences (Techniques) 

Novosibirsk

a-kugaevskikh@yandex.ruИнститут автоматики и электрометрии Сибирского отделения Российской академии наук; Новосибирский государственный университетРоссияInstitute of Automation and Electrometry of the Siberian Branch of the Russian Academy of Sciences; Novosibirsk State UniversityRussian Federation2021171120211935060Copyright © Кугаевских А.В., 20212021Кугаевских А.В.Kugaevskikh A.V.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://intechngu.elpub.ru/jour/article/view/173

Представлена модель комплексного нейрона конца линий. Этот тип нейронов дает максимальный отклик на конце линии и используется для уточнения выделения краев. В работе приведен обзор различных моделей нейронов конца линий и проведено сравнение их откликов. Нами предложена более простая и, в тоже время, точная CE-модель нейрона конца линий, базирующаяся на использовании фильтров Габора, находящихся в противофазе. Предложенная модель интегрирована в сверточную нейронную сеть выделения краев, описание которой также приведено в статье. Предлагаемая модель показала высокую эффективность.

This article is dedicated to modeling the end-stopped neuron. This type of neuron gives the maximum response at the end of the line and is used to refine the edge. The article provides an overview of different models of end-stopped neurons. I have proposed a simpler and more accurate model of an end-stopped neuron based on the use of Gabor filters in antiphase. For this purpose, the models of simple and complex cells whose output is used in the proposed model are also described. Simple cells are based on the use of a Gabor filter, the parameters of which are also described in this article. The proposed model has shown its effectiveness.

нейрон конца линийвыделение краевнейронные сетифильтр Габорапростые клеткикомплексные клеткиend-stopped neuronedge detectionneural networkGabor filtersimple cellcomplex cellИсследование выполнено при финансовой поддержке РФФИ в рамках научного проекта № 18-37-00029This paper was financially supported by the Russian Foundation for Basic Research (grant no. 18-37-00029)ReferencesSerre T., Wolf L., Bileschi S., Riesenhuber M., Poggio T. Robust Object Recognition with Cortex-Like Mechanisms. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, vol. 29, no. 3, pp. 411–426. DOI 10.1109/TPAMI.2007.56Serre T., Wolf L., Bileschi S., Riesenhuber M., Poggio T. Robust Object Recognition with Cortex-Like Mechanisms. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, vol. 29, no. 3, pp. 411–426. DOI 10.1109/TPAMI.2007.56Díaz-Pernas F. J., Martínez-Zarzuela M., Antón-Rodríguez M., González-Ortega D. Learning and surface boundary feedbacks for colour natural scene perception. Applied Soft Computing, 2017, vol. 61, pp. 30–41. DOI 10.1016/j.asoc.2017.07.055Díaz-Pernas F. J., Martínez-Zarzuela M., Antón-Rodríguez M., González-Ortega D. Learning and surface boundary feedbacks for colour natural scene perception. Applied Soft Computing, 2017, vol. 61, pp. 30–41. DOI 10.1016/j.asoc.2017.07.055Hubel D. H. Eye, brain, and vision. New York, Scientific American Library: Distributed by W. H. Freeman, 1988.Hubel D. H. Eye, brain, and vision. New York, Scientific American Library: Distributed by W. H. Freeman, 1988.Yazdanbakhsh A., Livingstone M. S. End stopping in V1 is sensitive to contrast. Nature Neuroscience, 2006, vol. 9, no. 5, pp. 697–702. DOI 10.1038/nn1693Yazdanbakhsh A., Livingstone M. S. End stopping in V1 is sensitive to contrast. Nature Neuroscience, 2006, vol. 9, no. 5, pp. 697–702. DOI 10.1038/nn1693Eskikand P. Z., Kameneva T., Ibbotson M. R., Burkitt A. N., Grayden D. B. A Possible Role for End-Stopped V1 Neurons in the Perception of Motion: A Computational Model. PLOS ONE, 2016, vol. 11, no. 10, p. e0164813. DOI 10.1371/journal.pone.0164813Eskikand P. Z., Kameneva T., Ibbotson M. R., Burkitt A. N., Grayden D. B. A Possible Role for End-Stopped V1 Neurons in the Perception of Motion: A Computational Model. PLOS ONE, 2016, vol. 11, no. 10, p. e0164813. DOI 10.1371/journal.pone.0164813Nicholls J. G. (ed.) From neuron to brain. 5th ed. Sunderland, Mass, Sinauer Associates, 2012.Nicholls J. G. (ed.) From neuron to brain. 5th ed. Sunderland, Mass, Sinauer Associates, 2012.Kugaevskikh A. V., Sogreshilin A. A. Analyzing the Efficiency of Segment Boundary Detection Using Neural Networks. Optoelectron. Instrument. Proc., 2019, vol. 55, no. 4, pp. 414–422. DOI 10.3103/S8756699019040137Kugaevskikh A. V., Sogreshilin A. A. Analyzing the Efficiency of Segment Boundary Detection Using Neural Networks. Optoelectron. Instrument. Proc., 2019, vol. 55, no. 4, pp. 414–422. DOI 10.3103/S8756699019040137Gabor D. Theory of communication. Journal of the Institution of Electrical Engineers, 1946, vol. 93, no. 26, pp. 429–457.Gabor D. Theory of communication. Journal of the Institution of Electrical Engineers, 1946, vol. 93, no. 26, pp. 429–457.Daugman J. G. Uncertainty relation for resolution in space, spatial frequency, and orientation optimized by two-dimensional visual cortical filters. J. Opt. Soc. Am. A, 1985, vol. 2, no. 7, p. 1160. DOI 10.1364/JOSAA.2.001160Daugman J. G. Uncertainty relation for resolution in space, spatial frequency, and orientation optimized by two-dimensional visual cortical filters. J. Opt. Soc. Am. A, 1985, vol. 2, no. 7, p. 1160. DOI 10.1364/JOSAA.2.001160Petkov N. Biologically motivated computationally intensive approaches to image pattern recognition. Future Generation Computer Systems, 1995, vol. 11, no. 4–5, pp. 451–465. DOI 10.1016/0167-739X(95)00015-KPetkov N. Biologically motivated computationally intensive approaches to image pattern recognition. Future Generation Computer Systems, 1995, vol. 11, no. 4–5, pp. 451–465. DOI 10.1016/0167-739X(95)00015-KKruizinga P., Petkov N. Nonlinear operator for oriented texture. IEEE Transactions on Image Processing, 1999, vol. 8, no. 10, pp. 1395–1407. DOI 10.1109/83.791965Kruizinga P., Petkov N. Nonlinear operator for oriented texture. IEEE Transactions on Image Processing, 1999, vol. 8, no. 10, pp. 1395–1407. DOI 10.1109/83.791965Bovik A. C., Clark M., Geisler W. S. Multichannel texture analysis using localized spatial filters. IEEE Trans. Pattern Anal. Machine Intell., 1990, vol. 12, no. 1, pp. 55–73. DOI 10.1109/34.41384Bovik A. C., Clark M., Geisler W. S. Multichannel texture analysis using localized spatial filters. IEEE Trans. Pattern Anal. Machine Intell., 1990, vol. 12, no. 1, pp. 55–73. DOI 10.1109/34.41384Pollen D. A., Ronner S. F. Visual cortical neurons as localized spatial frequency filters. IEEE Trans. Syst., Man, Cybern., 1983, vol. SMC-13, no. 5, pp. 907–916. DOI 10.1109/TSMC. 1983.6313086Pollen D. A., Ronner S. F. Visual cortical neurons as localized spatial frequency filters. IEEE Trans. Syst., Man, Cybern., 1983, vol. SMC-13, no. 5, pp. 907–916. DOI 10.1109/TSMC. 1983.6313086Akbarinia A., Parraga C. A. Feedback and Surround Modulated Boundary Detection. International Journal of Computer Vision, 2018, vol. 126, no. 12, pp. 1367–1380. DOI 10.1007/s11263-017-1035-5Akbarinia A., Parraga C. A. Feedback and Surround Modulated Boundary Detection. International Journal of Computer Vision, 2018, vol. 126, no. 12, pp. 1367–1380. DOI 10.1007/s11263-017-1035-5Bhattacharjee S. K., Vandergheynst P. End-stopped wavelets for detecting low-level features. In: SPIE’s International Symposium on Optical Sience, Engineering, and Instrumentation. Denver, CO, Oct. 1999, pp. 732–741. DOI 10.1117/12.366829Bhattacharjee S. K., Vandergheynst P. End-stopped wavelets for detecting low-level features. In: SPIE’s International Symposium on Optical Sience, Engineering, and Instrumentation. Denver, CO, Oct. 1999, pp. 732–741. DOI 10.1117/12.366829Rodrigues J., J. Buf M. H. du. Multi-scale keypoints in V1 and beyond: Object segregation, scale selection, saliency maps and face detection. Biosystems, 2006, vol. 86, no. 1–3, pp. 75–90. DOI 10.1016/j.biosystems.2006.02.019Rodrigues J., J. Buf M. H. du. Multi-scale keypoints in V1 and beyond: Object segregation, scale selection, saliency maps and face detection. Biosystems, 2006, vol. 86, no. 1–3, pp. 75–90. DOI 10.1016/j.biosystems.2006.02.019Würtz R. P., Lourens T. Corner detection in color images through a multiscale combination of end-stopped cortical cells. Image and Vision Computing, 2000, vol. 18, no. 6–7, pp. 531–541. DOI 10.1016/S0262-8856(99)00061-XWürtz R. P., Lourens T. Corner detection in color images through a multiscale combination of end-stopped cortical cells. Image and Vision Computing, 2000, vol. 18, no. 6–7, pp. 531–541. DOI 10.1016/S0262-8856(99)00061-XHeitger F., Rosenthaler L., Heydt R. von der, Peterhans E., Kübler O. Simulation of neural contour mechanisms: from simple to end-stopped cells. Vision Research, 1992, vol. 32, no. 5, pp. 963–981. DOI 10.1016/0042-6989(92)90039-LHeitger F., Rosenthaler L., Heydt R. von der, Peterhans E., Kübler O. Simulation of neural contour mechanisms: from simple to end-stopped cells. Vision Research, 1992, vol. 32, no. 5, pp. 963–981. DOI 10.1016/0042-6989(92)90039-LFukushima K. Neocognitron for handwritten digit recognition. Neurocomputing, 2003, vol. 51, pp. 161–180. DOI 10.1016/S0925-2312(02)00614-8Fukushima K. Neocognitron for handwritten digit recognition. Neurocomputing, 2003, vol. 51, pp. 161–180. DOI 10.1016/S0925-2312(02)00614-8

The authors declare that there are no conflicts of interest present.