The Research And Application Of Space Curve Matching Algorithm

Space curve matching is one of the fundamental issues in computer vision and computer graphics, with myriad applications ranging from3D shape retrieval to art restoration to medical diagnosis. The difficulty of the problem is how to detect the feature of curves accurately. The studying on2D curve matching has made some progress using V system. Furthermore, it generalized into matching two space curves which are similar. This paper is composed as follows:Firstly, this paper elaborates the relationship between shape matching and curve matching and introduces the current summary of the research on them.Secondly, we review the state of the art in shape matching methods which include shape matching based feature and invariants as well as descriptors. Then, we introduce the space curve matching processes and surface contour extraction method.Finally, a new3D space curve matching algorithm is proposed based on the V-system theory. The main idea is to describe space curves by V system so that their V descriptors can be obtained. According to the distance between their normalized V descriptors the matching curves will be found, and then the matching parameters are evaluated. Compared to the plane curve matching algorithm based on V system, it is difficult to estimate rotation matrix. We settled down the difficulty by using for reference of pose estimation algorithm in computer vision. Finally, a lot of simulation and real tests show that our algorithm is feasible and efficient.