(20) * << * >> * Russian * English * Content * All Issues

Comparison of standard and object by contour deviation

N.L. Kazanskiy, R.V. Khmelev

Image Processing Systems Institute of RAS

 PDF, 323 kB

Pages: 128-133.

Full text of article: Russian language.

Abstract:
It is known that an object shape is determined by its contours, and many objects can be recognized by their contours alone, without using the original halftone images. At present, numerous algorithms are available that allow to perform object recognition based on the contours. This article presents an approach with the following features: firstly, the exact matching of pixels of the standard and the object is not significant for this algorithm, as well as the discontinuity of the contours, and secondly, the calculated numerical characteristics are directly and unambiguously related to the shape of the objects.

Citation:
Kazanskiy NL, Khmelev RV. Comparison of standard and object by contour deviation. Computer Optics 2000; 20: 128 - 133.

References:

  1. Duda RO, Hart PE. Pattern classification and scene analysis. Moscow: Mir Publisher; 1976: 512.
  2. Tu J, Gonzalez R. Principles of pattern recognition. Moscow: Mir Publisher; 1978; 412.
  3. Glumov NI. Development and application of moment invariants for image processing in a sliding window. Computer Optics 1995; 14-15(1): 46-54,
  4. Kazanskiy NL, Myasnikov VV, Khmelev RV. Algorithms for computing distances to object pixels on binary images. Computer Optics No. 20, 2000. p. 134-139.

© 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