Research On Shape Description And Matching

2D shape description and matching is a key issue in the field of computer vision and pattern recognition. Shape description and matching based on the contour sample points is a mainstream technology. But the richness of shape descriptor and the matching of shapes with narrow gap still can not meet the application requirements. This paper implements a 2D shape description and matching method based on vector function, innovative work as follows:We propose shape description and matching algorithm based on the vector function. The algorithm uses a series of vectors between the contour sample points to build shape descriptors, and describe the shape characteristic in continuous domain, which combines the advantages of Shape Context and Height Function. Experimental results show that the algorithm can be obtained higher shape retrieval accuracy.We proposed slit shape matching method based filtering algorithm. The method is for slit shape uncorrent matching problem, the key of slit shape recognition is to mearsure the difference betweent the shape before morphological process and the shape after morphological process. Experimental results show that the algorithm can effectively filter out the shape with slit structures, improving shape retrieval accuracy while maintaining the shape retrieval efficiency, and solve the problem of the uncorrect matching of slit shapes.