电力电子第六章.pptVIP

6.4.4 Graph and Tree Matching Consider two graphs (or trees) G1=﹛N1, R1﹜and G2=﹛N2, R2﹜, describing the structural relations of two patterns. A homomorphism from GI to G2 is any function f establishing a correspondence between the nodes of N1 and the nodes of N2 such that any pair of adjacent nodes in GI corresponds, through f, to adjacent nodes in N2. Let (a, b) represent two adjacent nodes of N1. Then a homomorphism f satisfies: (6-22) Figure 6.15. Two digraphs with two homomorphisms: (a) f( l)=a, f(2)=b, f(3)=c; (b) f( l)=a, f(2)=c, f(3)=d. Figure 6.16. Match graph for the two graphs shown in Figure6.15.The homomorphiss are shown with solid line. * * Pattern RecognitionChapter 6 Structural Pattern Recognition (Pattern Recognition Concepts, Methods and Applications J.P. Marques de Sa Springer) 6 Structural Pattern Recognition 6.1 Pattern Primitives 6.2 Structural Representations 6.3 Syntactic Analysis 6.4 Structural Matching 6.1 Pattern Primitives 6.1.1 Signal Primitives a piecewise linear approximation of a signal is by the most popular method of signal decomposition.Set a signal s(x) that the approximate by a piecewise linear function h(x) with d segments. The approximation error is: （6-1） where an appropriate norm, usually the Chebychev norm or the Euclidian norm, is used to evaluate the deviations of s(xj) from hi(xj). 6.1.2 Image Primitives Image primitives can be obtained through the application of a large variety of image analysis techniques, such as image segmentation, edge detection, contour following and medial axis transformation. Chain Code and Templates Figure 6.1 (a) Octal primitives (top) and templates (bottom); (b) Binary image with contour line seg