(45-4) 12 * << * >> * Russian * English * Content * All Issues

Investigation of the applicability of the convolutional neural network U-Net to a problem of segmentation of aircraft images
D.A. Gavrilov 1,2

1 Lebedev Institute of Precise Mechanics and Computer Engineering, Russian Academy of Sciences,
Russian Federation, Moscow, 51, Leninskiy boulevard, 119991,
2 Moscow Institute of Physics and Technology,
Russian Federation, 9 Institutskiy per., Dolgoprudny, Moscow Region, 141701

 PDF, 1365 kB

DOI: 10.18287/2412-6179-CO-804

Pages: 575-579.

Full text of article: Russian language.

Abstract:
The paper investigates the applicability of the convolutional neural network "U-Net" to a problem of segmentation of aircraft images. The neural network image segmentation method is based on the "Carvana" implementation with the "U-Net" architecture. For orientation recognition, a neural network built in the Keras open neural network library based on the pretrained VGG16 neural network is used. The approach considered allows the image segmentation to be conducted. The results of the experiments have shown the possibility of a fairly accurate selection of the object of interest. The resulting binary masks make it possible to visually classify the aircraft in the image.

Keywords:
technical vision, detection, localization, neural network, recognition, image processing.

Citation:
Gavrilov DA. Investigation of the applicability of the convolutional neural network U-Net to a problem of segmentation of aircraft images. Computer Optics 2021; 45(4): 575-579. DOI: 10.18287/2412-6179-CO-804.

References:
  1. Belov AM, Denisova AY. Scene distortion detection algorithm using multitemporal remote sensing images. Computer Optics 2019; 43(5): 869-885. DOI: 10.18287/2412-6179-2019-43-5-869-885.
  2. Borzov SM, Guryanov MA, Potaturkin OI. Study of the classification efficiency of difficult-to-distinguish vegetation types using hyperspectral data. Computer Optics 2019; 43(3): 464-473. DOI: 10.18287/2412-6179-2019-43-3-464-473.
  3. Lovtsov DA, Gavrilov DA. Automated special purpose optical electronic system’s functional diagnosis. International Seminar on Electron Devices Design and Production (SED) 2019: 8798409.
  4. Gavrilov DA. Quality assessment of objects detection and localization in а video stream [In Russian]. Herald of the Bauman Moscow State Technical University, Series Instrument Engineering 2019; 125(2): 40-55.
  5. Gavrilov DA, Pavlov AV. Streaming hardware based implementation of SURF algorithm [In Russian]. Proceedings of Universities. Electronics 2018; 23(5): 502-511.
  6. Parkalov AV. Application of neural and semantic networks in the segmentation of the earth's surface bitmaps [In Russian]. Open Semantic Technologies for Intelligent Systems (OSTIS-2012) 2012: 527-530.
  7. Revyakin AM, Skurnovich AV. Approaches to the development of a recognition system to solve the problem of determining the content of digital images [In Russian]. Naukovedenie 2016; 8(4). Source: <https://naukovedenie.ru/PDF/30TVN416.pdf>.
  8. Dyudin MV, Povalyaev AD, Podvalny ES, Tomakova RA. Methods and algorithms of contour analysis for classification problems of complexly structured images [In Russian]. Bulletin of Voronezh State Technical University 2014; 3(1): 54-59.
  9. Mestetsky LM. Continuous morphology of binary images: figures, skeletons, circulars [In Russian]. Moscow: "Fizmatlit" Publisher; 2009.
  10. Druki AA, Spitsyn VG, Boltova YuA, Bashlykov AA. Sematic segmentation of earth remote sensing data using neural network algorithms [In Russian]. Bulletin of the Tomsk Polytechnic University. Series Engineering of Georesources 2018; 329(1): 59-68.
  11. Ronneberger O., Fischer Ph., Brox Th. U-net: Convolutional networks for biomedical image segmentation. In Book: Navab N, Hornegger J, Wells WM, Frangi AF, eds. Medical image computing and computer-assisted intervention – MICCAI 2015. Pt III. New York: Springer International Publishing Switzerland; 2015: 234-241. DOI: 10.1007/978-3-319-24574-4_28.
  12. Carvana image masking challenge – 1st place winner’s interview. Source: <http://blog.kaggle.com/2017/12/22/carvana-image-masking-first-place-interview/>.
  13. Keras. Source: <https://keras.io/>.
  14. Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. ICLR-2015 2015: 1-14.
  15. Nanda Y. What is the VGG neural network? Source: <https://www.quora.com/What-is-the-VGG-neural-network>.
  16. Gavrilov DA, Mestetskiy LM, Semenov AB. A method for aircraft labeling in remote sensing images based on continuous morphological models. Program Comput Softw 2019; 45(6): 303-310.

© 2009, IPSI RAS
151, Molodogvardeiskaya str., Samara, 443001, Russia; E-mail: journal@computeroptics.ru ; Tel: +7 (846) 242-41-24 (Executive secretary), +7 (846) 332-56-22 (Issuing editor), Fax: +7 (846) 332-56-20