关键词: 模式识别, 形状描述子, 形状匹配, 不变量

Abstract: Through the research about invariants under similarity transformations（rotation,scale and translation）,the proof of the fact that the center of gravity of an image and two endpoints of its major axis are such invariants under similarity transformations is stated,and a geometric feature-based shape descriptor θ invariant to rotation,scale and translation is proposed in this paper.θ is the included angle between the line segments connecting center of gravity and two endpoints of major axis.The theoretical analysis and experiments both show that the shape descriptor is efficient and simple,and gains desirable matching results.

Key words: pattern recognition, shape descriptor, shape matching, invariant