Fast and robust object recognition based on reduced Shock-Graphs and Pruned-SIFT

Abstract:

Fast algorithms for image recogition have become a priority when the number of images to be analyzed can be counted in the number of millions. Therefore, the need for fast algorithms for image recognition. This paper describes an algorithm capable of recognize an object in an image by the fusing information from a reduced version of the Shock-Graph algorithm and the Scale Invariant Feature Transform (SIFT) points. The proposed algorithm uses the reduced Shock-Graph to obtain a skeleton of an object in order to minimize the number of SIFT points to reduce the computational complexity of object image comparison. The proposed algorithm is capable of recognizing objects in a fast way under rotation, deformation and scaling. Using a collection of shapes, we demonstrate the performance of our implementation using a combination of the reduced Shock-Graph and SIFT points. © Springer-Verlag 2013.

Año de publicación:

2013

Keywords:

Fuente:

scopus