PRECISION, SPEED AND COMPLEXITY OF DEVICES FOR IMAGE CODING BY CONTROL POINTS

• Main Authors: M. M. Almahrouq, A. I. Bobrovsky, M. M. Eid, Y. M. Sokolov, A. Salem, S. S. Fahmi
• Format: Article
• Language: English
• Published: Saint Petersburg National Research University of Information Technologies, Mechanics and Optics (ITMO University) 2016-07-01
• Series: Naučno-tehničeskij Vestnik Informacionnyh Tehnologij, Mehaniki i Optiki
• Subjects: recursion; polygon; control points; encoding; precision; speed and complexity
• Online Access: http://ntv.ifmo.ru/file/article/15813.pdf

Subject of Research. We have proposed a method, algorithms and devices for image coding and decoding by control points without switching to the spectral domain of the signal in order to determine information quality indicators of image processing systems. Formalized consideration of the coding devices complexity stimulated by the creation and development of VLSI and systems on a chip requires revision of "efficient" encoding concept, since not only the transmission accuracy (error) and the transmission speed, but also complexity enters the circle of considered variables taken into account at the source encoding. Method. The proposed approach is based on the application of: firstly, spatial-recursive method of partitioning the images into polygons of various shape and size when searching the reference feature points of the images; secondly, regular and irregular triangulation algorithms at the recovery phase of the resulting image. Main Results. As a result of the simulation of image coding and decoding algorithms based on control points we have obtained three-dimensional graphic comparison of data quality indicators: error, transmission speed and complexity of devices for coding and decoding by control points with the other known image coding algorithms. Practical Relevance. The proposed algorithms for images partition and search of control points along with low significance of difficulty give the possibility: firstly, to reduce the transmission rate by 1.5-2 times as compared to standard algorithms for spectral conversion; secondly, to have a compact representation of video information in the form of one-dimensional dynamic array of control points with the possibility of semantic analysis and the transition from standard formats for video data transmission and storage to specific forms, focused on solving specific problems.