MODIFICATION OF THE ALGORITHM OF THE CARCASS OF ZONG-SUN FOR THE PROBLEM OF HANDWRITING RECOGNITION

The problem of recognizing the shape of objects that can be represented as images is considered in many applications. To implement vectorization algorithms, refinement algorithms must be applied, the result of their work is a reduction in the number of pixels that need to be processed, while information about the shape and topology of the analyzed image is stored. This paper analyzes the algorithms of skeletization-converting images of symbols into thin lines, which operate either by boundary vectors or pixels. It is noted that they all build lines with a thickness of more than one pixel, although it can even achieve a gain in the speed of calculations. This leads to the fact that the construction of the graph skeleton will appear unnecessary nodes. In most methods of thinning preserves the connectivity. The author proposes an algorithm that improves The Zong-Sun algorithm, which in the construction of the image graph does not lead to the fact that there are extra branching points. Unlike the proposed algorithm, the template method and the Zong-Sun algorithm do not give a line with a thickness of one pixel.

1. Matematicheskaya morfologiya. Fil'tratsiya [Elektron. resurs]. – Rezhim dostupa: http://graphics.cs.msu.su/courses/cg02b/lectures/lection4/, svobodnyy. – Zagl. s ekrana.

2. Lyu Wenyin From Raster to Vectors: Extracting Visual Information from Line Drawings / Lyu Wenyin, DovDori // [Elektronnyy resurs]: – Rezhim dostupa: http://research.microsoft.com/research/pubs/view.aspx?pubid=924, svobodnyy.

3. Novikov Yu.L. Issledovanie metodov vektorizatsii rastrovykh izobrazheniy i ikh realizatsiya v geoinformatsionnoy sisteme / Yu.L. Novikov // Tomsk: Tomsk. gos. un-t. Fakul'tet informatiki, 1999.-177 p.

4. Bukhshtaber V. M. Avtomatizirovannaya sistema analiza ploskikh tochechnykh izobrazheniy metodom skeletizatsii kak instrument resheniya zadach prikladnoy statistiki / V. M. Bukhshtaber, V. M.Klyatskin, V. V.Mottl', E.V.Shchepin // Programmnye produkty i sistemy, 3 (1991), pp. 52–62

5. Kushnir O.A. Sravnenie formy binarnykh rastrovykh izobrazheniy na osnove skeletizatsii / O.A.Kushnir // Mashinnoe obuchenie i analiz dannykh, 2012. Vol. 1, No. 3. pp. 252-263.