THE ALGORITHM FOR CONTOUR SEGMENTATION OF IMAGES WITH HIERARCHICAL OBJECTS ON THE BASIS OF THE CANNY METHOD WITH USE OF WAVELET-TRANSFORMATION

Abstract

The existing methods for contour segmentation don’t take into consideration the hierarchical structure of the objects what leads to decrease in efficiency of a pattern recognition system. The purpose of this work is to develop an algorithm for contour segmentation of images with hierarchical structured objects in order to improve the quality of segmentation on the lower values of SNR and to increase the efficiency of further work with the received edges in the pattern recognition system. The study results can help increase the efficiency in such a way. The first length of the wavelet being set must be bigger than the length of each of the objects that are desired to be ignored in the edge. After applying the proposed algorithm, we obtain the edge of the large object only.  If then the system requires internal, smaller parts for further recognition process, an edge tracking must be made and after choosing the new length of the wavelet again the stages of the proposed algorithm should be applied – now on the image limited by the traced edge. As a result, the system obtains the edges of the desired object instead of getting all the unwanted object details and searching for the desired one among them. Canny edge detection method was chosen as the basis for the algorithm being developed. It was proposed to replace its first two steps with a wavelet-transformation, what allowed to develop the algorithm, which provides high noise immunity and adjustable level of detail.

Author Biographies

Віктор Миколайович Крилов, Odessa National Polytechnic University

Dr. of Science, Professor of the Department of applied mathematics
and information technologies, Odessa National Polytechnic University

Тетяна Пардіпівна Кумар, Odessa National Polytechnic University

Student of the Department of applied mathematics and information technologies, Odessa National Polytechnic University

Published
2017-12-28
How to Cite
Крилов, В., & Кумар, Т. (2017). THE ALGORITHM FOR CONTOUR SEGMENTATION OF IMAGES WITH HIERARCHICAL OBJECTS ON THE BASIS OF THE CANNY METHOD WITH USE OF WAVELET-TRANSFORMATION. Electrotechnic and Computer Systems, (26(102), 88-93. https://doi.org/10.15276/eltecs.26.102.2017.10
Section
Artificial Intellect Systems