Shape Matching Using Metric Partition Constraint

A novel shape matching method is proposed based on metric partition constraint.The method can obtain high recognition rate with low time complexity, and is also robust to geometric deformations.First of all, the metric information, such as Euclidean distance and triangle radius, is extracted for each sample point, which can constrain the geometric relations between all sample points and the whole shape.Furthermore, the outer contour of the shape is divided into several partitions, and smoothed separately to make it robust to deformations.Finally, dynamic programming algorithm is employed to compare different shapes.Experiments on public databases show that the proposed method is effective and fast, and is also adapt to various deformations.It is worth mentioning that the proposed method can be extended with numerous metrics.