New dominant point detection for image recognition

In this paper, a new dominant point detection scheme is introduced. The previous approaches for detecting dominant points mainly rely on the method of searching for the relationship among points under consideration and their neighboring points. However, these approaches usually require much computing power. Therefore, this motivates us to develop a new method for detecting dominant points, in which less computation is required. Our method which is a global approach is based on the detection of the zero crossing point of the differences in distances and angular differences of three adjacent points on the curve from a fixed point. Experiment results show that our approach can detect dominant points with great satisfaction and fast speed.