文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 高等教育 > 理学 > A Survey of Shape Feature Extraction Techniques
侵权投诉

A Survey of Shape Feature Extraction Techniques

A Survey of Shape Feature Extraction Techniques

56 Pattern Recognition Techniques, Technology and Applications When the contour is traversed in counter clockwise direction, positive, negative and zero values of TAR mean convex, concave and straight-line points, respectively. Figure 11 demonstrates these three types of the triangle areas and the complete TAR signature for the hammer shape.

A Survey of Shape Feature Extraction Techniques

Fig. 11. Three different types of the triangle-area values and the TAR signature for the hammer shape

By increasing the length of the triangle sides, i.e., considering farther points, the function of Eq. 16 will represent longer variations along the contour. The TARs with different triangle sides can be regarded as different scale space functions. The total TARs, ts ∈ [1, N/2 - 1], compose a multi-scale space TAR.

In [21], authors show that the multi-scale space TAR is relatively invariant to the affine transform and robust to non-rigid transform.

3.7 Chord length function

The chord length function is derived from shape boundary without using any reference point. For each boundary point P, its chord length function is the shortest distance between P and another boundary point P’ such that line PP’ is perpendicular to the tangent vector at P [10]. The chord length function is invariant to translation and it overcomes the biased reference point (which means the centroid is often biased by boundary noise or defections) problems. However, it is very sensitive to noise, so that there may be drastic burst in the signature of even smoothed shape boundary.

3.8 Discussions

A shape signature represents a shape by a 1-D function derived from shape contour. To obtain the translation invariant property, they are usually defined by relative values. To obtain the scale invariant property, normalization is necessary. In order to compensate for orientation changes, shift matching is needed to find the best matching between two shapes. Having regard to occultation, Tangent angle, Contour curvature and Triangle-area representation have invariance property. In addition, shape signatures are computationally simple.

Shape signatures are sensitive to noise, and slight changes in the boundary can cause large errors in matching procedure. Therefore, it is undesirable to directly describe shape using a shape signature. Further processing is necessary to increase its robustness and reduce the matching load. For example, a shape signature can be simplified by quantizing the signature into a signature histogram, which is rotationally invariant.

上一页第14页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email