Speaker
Prof.
Ivan Gostev
(NRU HSE)
Description
Process of recognition of the shape of graphic objects consists of several stages. At the first stage as a result of processing images some set of characteristic properties of some object are extracted. On the second one, identification of object is made by comparison of these properties with properties of the sample. Presence of noise on real graphical images often distorts quality of characteristic properties. This article describes methods for the extraction of characteristic properties of graphical objects and presents methods for the identification of object’s shape with partly-present or distorted characteristic properties. The advantage of the described methods is their invariance to affine transformations of the shape of object, and also high speed of identification independent on complexity of the object being identified.
Primary author
Prof.
Ivan Gostev
(NRU HSE)
Co-author
Prof.
Leonid Sevastyanov
(PFUR)