Detection of dominant points based on noise suppression and error minimisation

摘要

Approximation of contours has been shown to be reliant upon accurate determination of the region of support. In this paper, it is proposed that an adaptive approach will provide improved results and better shape representation. Local optimisation of chords along with desensitising the arc-chord measure to small perturbations is used. Experiment results show that the proposed method is superior to existing methods.