Curve length estimation using vertix chain code

Abstract

Most of the applications in image analysis are based on Freeman chain code. In this paper, for the first time, vertex chain code (VCC) proposed by Bribiesca is applied to improve length estimation of the 2D digitized curve. The chain code has some preferences such as stable in shifting, turning, mirroring movement of image and has normalized starting point. Due to the variety of length estimator methods, we focused on the three specific techniques. First, the way Bribiesca proposed which is based on counting links between vertices; second, based on maximum length digital straight segments (DSSs) and lastly local metrics. The results of these length estimators with the real perimeter are compared. Results thus obtained exhibits that length estimation using VCC is nearest to the actual length.