RUS  ENG
Full version
JOURNALS // Vestnik Moskovskogo Universiteta. Seriya 1. Matematika. Mekhanika // Archive

Vestnik Moskov. Univ. Ser. 1. Mat. Mekh., 2018 Number 6, Pages 20–30 (Mi vmumm580)

This article is cited in 2 papers

Mathematics

Implementation of the fast algorithm for geometrical coding of digital images with the use of CUDA architecture

A. Yu. Chekunov

Lomonosov Moscow State University, Faculty of Mechanics and Mathematics

Abstract: In this paper we compare the speed and quality of CUDA realization for new contour detection algorithm based on geometrical coding with CUDA implementation of Canny algorithm which is commonly used in OpenCV library. The comparison shows that the new approach can really compete with the Canny operator and in some cases even overcome it in speed and quality. Examples of geometrical coding contour detection in different situations are presented.

Key words: pattern recognition, geometrical coding, coding surface, contour analysis, edge detection, computer vision, image processing, image stitching.

UDC: 511

Received: 14.02.2018


 English version:
Moscow University Mathematics Bulletin, Moscow University Måchanics Bulletin, 2018, 73:6, 229–238

Bibliographic databases:


© Steklov Math. Inst. of RAS, 2024