一种基于轮廓线的形状描述与匹配方法
doi: 10.3724/SP.J.1146.2006.01468
A Contour-Based Shape Description and Matching Method
-
摘要: 该文提出了一种新的基于目标轮廓线的形状描述与匹配方法。该方法将两个待匹配的目标的轮廓线等弧长地分割成相同数目的弧段，用直线段连接相邻的割点构成对轮廓线的多边形近似，用多边形的顶点到形状的几何中心的距离和其两个后继顶点的距离构成的3个距离串来描述形状，形状之间的相似度用它们的3个距离串的欧氏距离和来计算。这种描述方法满足惟一性、紧致性和不变性，而且计算简单，对形状的全局特征和局部特征都能很好地描述。实验结果表明，将该方法用于形状匹配具有较高的精度和可靠性。Abstract: A novel contour-based shape description and matching method is proposed. It equally segments the contours of the two objects into same numbers of arcs, an approximating polygon is then constructed with connecting the neighbor segmenting points using line segments. The three distance strings which are formed with the distance from the vertexes to the center of shape and the distances from the vertexes to the two subsequence vertexes are used to describe the shape, and the sum of the distances between their three distance strings is used to measure the difference between them. This method possesses uniqueness, compactness and invariance. Furthermore, this method is very simple to compute and can reflect both the global feature and the local feature of the shape. Experiment results show that the proposed method is accurate and reliable for matching shapes.
-
Gonzalez R C and Woods R E. Digital Image Processing[M]. America, Pearson Education, Inc., publishing as PrenticeHall. Second Edition, 2002: 648-649.[2]Morkhrarian F S. Silhouette-based occluded objectrecognition through curvature scale space [J].Mach. VisionAppl.1997, 10(3):87-97[3]Bernier T and Landry J-A. A new method for representingand matching shapes of natural objects [J]. PatternRecognition, 2003, 36(8): 1711-1723.[4]Kindratenko V V. On using functions to describe the shape[J].Journal of Mathematical Imaging and Vision.2003, 18(3):225-245[5]Tsai W H and Yu S S. Attributed string matching withmerging for shape recognition [J].IEEE Trans. on PatternAnalysis and Machine Intelligence.1985, 7(4):453-462[6]Groskey W I and Mehrotra R. Index-based object recognitionin pictorial data management [J].Comput. Vision GraphicsImage Process.1990, 52(3):416-436[7]赵宇, 陈雁秋. 曲线描述的一种方法: 夹角链码 [J]. 软件学报, 2004, 15(2): 300-307.[8]Sebastian T B, Klein P N, and Kimia B B. Recognition ofshapes by editing shock graphs[C]. Eighth IEEE InternationalConference on Computer Vision, Vancouver, Canada, 2001, 1:755-762. -
计量
- 文章访问数: 3075
- HTML全文浏览量: 78
- PDF下载量: 2818
- 被引次数: 0
下载:
