Detection of the main symmetry axis based on the clustering analysis

Abstract

In order to improve the detection performance of the main symmetry axis, a new detection method based on the clustering analysis is proposed. Some feature points are extracted from the image based on the Harris corner detection, and C-means clustering method is used to classify the feature points into C groups. Multiple random sampling is performed in each cluster to get the feature point pairs. The candidate symmetry axis can be determined by resolving the perpendicular bisector of each pair of feature points. The main symmetry axis can be determined according to the distribution of the candidate symmetry axes. Main symmetry axes of many images are detected by the method. Simulation results prove the method valid.