WLSF-method for build-up of feature description for image feature points neighborhoods

D. A. Pimankin
Nizhny Novgorod State Technical University n.a. R.E. Alekseev

 

Received November 6, 2012

 

Abstract.  Feature description build-up method for feature points neighborhoods based on image gradient vectors orientation analysis by weighted least squares (WLSF-method) is proposed. It is shown, that two modifications of this approach can be successfully applied for the task of feature points matching for a large class of images. The advantage of the WLSF-method over SURF-method in the case of noisy image analysis is shown.

Key words:  Weighted least squares, orthogonal transformation, feature space, feature point matching.