计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (1): 184-188.

• 图形图像处理 • 上一篇    下一篇

基于弦长位置矩阵的目标识别

柴  华1,杨明强2   

  1. 1.国家知识产权局专利局 专利审查协作北京中心,北京 100190
    2.山东大学 信息科学与工程学院,济南 250100
  • 出版日期:2015-01-01 发布日期:2015-01-06

Object recognition based on chord location matrix

CHAI hua1, YANG Mingqiang2   

  1. 1.Patent Examination Cooperation Center of the Patent Office, SIPO, Beijing 100190, China
    2.College of information Science and Engineering, Shandong University of Science and Technology, Jinan 250100, China
  • Online:2015-01-01 Published:2015-01-06

摘要: 为了提取具有仿射不变性的形状特征,提出一种具有仿射不变性的弦长位置矩阵算法。该算法将各个方向的弦长的位置以及同一位置上不同长度的弦长排序信息作为特征量,以构造弦长位置矩阵,作为目标形状的特征信息,并利用动态空间弯折距离作为测度距离。实验结果表明,应用提出的弦长位置矩阵作为形状特征对于仿射变换具有鲁棒性。

关键词: 弦长位置, 仿射不变, 特征提取

Abstract: To improve the affine invariance of features of shape, the chord location matrix algorithm is proposed. The chords location can be determined by affine invariance ratio of the closed areas. The chords location in all directions and sorting information of chords with different length on same position can construct chord location matrix as the feature vectors. The Dynamic Space Warping(DSW) algorithm proposed is used as similarity measure. Experimental results show that the approach is robust to affine transformation.

Key words: chord location, affine invariance, feature extraction