Make your own free website on Tripod.com
The principles of geometry of real lines. Interference defended approximation. Subdivision into uniform segments. Comparison of configurations.

Shulga V.I.

Ph.D., senior scientist, National Academy of Sciences of Ukraine (Kiev).

--------------------------------------------------------------------------------------------------------

It's addressed to persons and organizations which have interest to shape analysis (recognition) in frame of image, picture, drawing, scheme, map, graph, trace and signal interpretation in general.

--------------------------------------------------------------------------------------------------------

It is proposed for further joint advancement and practical adaption:

THE PROCEDURAL SEQUENCE FOR THE SEGMENTATION - APPROXIMATION AND INTEGRAL MATCHING OF THE LINE FORMED (CONTOUR) SHAPES.

That is based on:

- the interference defended "gentle extracting" approximation of line-like graphic patterns;

- the universal - "soft" subdivision of such patterns into separate - uniform segments;

- the fast integral (pose) identification of geometric shapes via "parametrical unison".

In whole, that provides the efficient manner for recognition afarbitrary formed (curvilinear) figures with respect to their samples in case when one figure may partially overlap another.

It may be used in computer, industrial, robot vision, - aerospace observation, navigation, object-target detection,- medical-biological microscopy, X-ray photography, nuclear track processing, drawings recognition, document analysis and others.

The developed procedure sequence has been applied to automatic visual pose identification of plate objects huddled on a surface. In future the announced method basis may be extended to solve the similar problems for 3D shapes. Besides the proposed concepts extend idea "arsenal" of shape analysis in general. The most complete description of this work were given in [1]. There is the brief review [2].

The work was approved by spesialists of Moscow State University, Institute of Applied Mathematics, Institute of Transmission Information Problems, Institute of Control Problems, Energy Institute and other organizations.

But the fatal collapse of the former USSR doesn't make it possible to continue and complete the work for practical use. First of all, it is needed to adapt the developed procedures into any modern program system. In addition technique of the computations may be essentially improved. This is start direction for further elaboration.

Have you interest and possibility to support and advance of the represented work? Please send your considerations and propositions.

Respectfully Vladimir I. Shulga

E-mail: shulga@altavista.net

Contact address: Shulga V.I. / d.27, korp.1, kv.12 / Chokolovsky Blvd.

/ Kiev-87 / Ukraine, 252087

Phones: +38-044-242-8053, +38-044-221-0753
 
 

1. Shulga V.I. "Complex of Interference Defended Procedures of Contour Boundary Approximation and Recognition of Objects on Contour Images (research work)"/ Glushkov Institute of Cybernetics, Academy of Sciences of the Ukraine, Kiev, 1992, 75p. (in Russian) / Deposited in All-Union Institute of Scientific and Engineering Information (Moscow), 04.01.92, N 12-B92.

2. Shulga V.I. "The Procedural Sequence for the Segmantation - Approxi- mation and Integral Comparison of the Contour Shapes" / Second Asian Conference on Computer Vision (ACCV'95, December 5-8, 1995, Singapor).

P.S. The [1] has the following CONTENTS. To organize the operative professional translation of this manuscript into regular English a moderate financial support is needed. Such support may be considered as one of start aspects of the further co-operation.

In general the present work goes back to the investigation which has been announced in: Shulga V.I. Interference Defended Recognition of Contour Image of Flat and 3D Objects in Problem of the Visual Sensing of the Manipulate Robots. Dissertation Review. Istitute of Cybernetics, Acad. of Sc. of UkSSR, 1988, Kiev, 16p.

Back to English

Back to Russian