A corner detection method using angle accumulation
Rafi Md. Najmus Sadat, Zinat Sayeeda, Md. Musfequs Salehin, Naurin Afrin · 2011
In this paper a Contour based corner detector that provides robust corner locations against different image transformations has been proposed. The presented corner detector is the combination of one of the most outperformed corner detectors named CPDA and the fast and efficient high curvature point detector IPAN99. The combination has been done in such a way that the weaknesses from both of the algorithms are eliminated and thus a promising method has come out. A comprehensive performance evaluation of the proposed detector has been performed by using two metrics Average Repeatability and Localization Error. The experimental results demonstrate the effectiveness of the proposed approach among seven other existing contour based corner detectors.