Journal of Xidian University ›› 2016, Vol. 43 ›› Issue (4): 45-50.doi: 10.3969/j.issn.1001-2400.2016.04.009

• Article • Previous Articles     Next Articles

Gaussian shape context descriptors

FENG Xiangwei;FENG Dazheng;HOU Ruili   

  1. (National Key Lab. of Radar Signal Processing, Xidian Univ., Xi'an  710071, China)
  • Received:2015-05-14 Online:2016-08-20 Published:2016-10-12

Abstract:

The shape context (SC) descriptors based on the Gaussian window (GSC) are proposed. The GSC descriptors can be used to match shapes. Encircling the reference point, the GSC sets up a group of Gaussian window functions. According to the Euclidean distance between the center points and the sampling points, the GSC describes the spatial relations between sampling points based on the shape descriptors calculated by Gaussian window functions by the use of as many sampling points as possible. More accurate correspondences between two shapes are determined by comparing the similarities of the shape descriptors. Experimental results show that the GSC can find more accurate correspondence relations and provide a good initial value. The proposed descriptors can also make the matching algorithm converge fast, while keeping a good anti-noise and robust performance compared with SC.

Key words: shape context, Gaussian window, shape descriptors, shape matching, robust