Adaptive scale method for representing and matching shapes

作者：Wang Ning^*; Zhou Jingli; Chen Jiazhong; Xiang Jinhai

来源：Journal of Huazhong University of Science and Technology (Natural Science Edition), 2010, 38(6): 71-74.

• Adaptive scale
• Angle of rotation
• Contour features
• Contour points
• Euclidean distance
• Feature vectors
• Green's theorem
• Main characteristics
• Novel methods
• Polar coordinates system
• Scale Factor
• Shape matching
• Shape representation
• Shape representation and matching
• Similarity distance
• Similarity measurements

A novel method for shape representation and matching is presented. Firstly, the coordinate of contour centroid is calculated as the pole according to Green's theorem. Then contour points are sampled at equal angle interval in polar coordinates system. The ratios of area that are enclosed by the radius of contour points are regard as feature vectors of contour. Finally, an improved normalized Euclidean distance is adopted to calculate the similarity distance between two feature vectors. The new method of shape representation needs less data to represent the contour feature and preserve the main characteristic. The new method of similarity measurement can obtain angle of rotation, scale factor and flip sign effectively, which is not affected by zoom, flip, and rotation. Experimental results show that the proposed method is feasible and accurate.

• 出版日期2010