Optimal Differentiators for the Algorithm of Orientation Estimation on the Basis of a Gradient Structure Tensor
I. S. Gruzman
Novosibirsk State Technical University
isgruzman@gmail.com
Keywords: optimal mask differentiating filters, gradient structure tensor
Pages: 3-11
Abstract
A method of constructing optimal mask differentiating filters minimizing the level of the root-mean-square or systematic error of estimation of the anisotropic structure orientation observed in the presence of additive white noise is proposed. Results of a comparative analysis of the accuracy of gradient algorithms are presented. It is demonstrated that the use of optimal differentiating filters substantially increases the accuracy of the orientation estimation algorithm.
|
