外文翻譯--一種用于非剛性點(diǎn)定位的新匹配算法（英文）

1、See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/222658656A New Point Matching Algorithm for Non-Rigid RegistrationArticle in Computer Vision and Image Under

2、standing · February 2003DOI: 10.1016/S1077-3142(03)00009-2 · Source: CiteSeerCITATIONS1,247READS2,1162 authors:Some of the authors of this publication are also working on these related projects:Consciousness Vi

3、ew projectRemote Sensing View projectHaili ChuiHologic15 PUBLICATIONS 2,168 CITATIONS SEE PROFILEAnand RangarajanUniversity of Florida251 PUBLICATIONS 7,555 CITATIONS SEE PROFILEAll content following this page wa

4、s uploaded by Anand Rangarajan on 24 April 2017.The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original documentand are linked to publications on Res

5、earchGate, letting you access and read them immediately.Keywords: Registration; Non-rigid mapping; Correspondence; Feature-based; Softassign; Thin-plate splines (TPS); Robust point matching (RPM); Linear assignment; Outl

6、ier rejection; Permutation matrix; Brain mapping1. IntroductionFeature-based registration problems frequently arise in the domains of computer vision and medical imaging. With the salient structures in two images represe

7、nted as compact geometrical entities (e.g., points, curves, and surfaces), the registration prob- lem is to find the optimum or a good suboptimal spatial transformation/mapping be- tween the two sets of features. The poi

8、nt feature, represented by feature location is the simplest form of feature. It often serves as the basis upon which other more so- phisticated representations (such as curves, surfaces) can be built. In this sense, it c

9、an also be regarded as the most fundamental of all features. However, feature-based registration using point features alone can be quite difficult. One common factor is the noise arising from the processes of image acqui

10、sition and feature extraction. The presence of noise means that the resulting feature points cannot be exactly matched. Another factor is the existence of outliers—many point features may exist in one point-set that have

11、 no corresponding points (homologies) in the other and hence need to be rejected during the matching process. Finally, the geometric transformations may need to incorporate high dimensional non-rigid mappings in order to

12、 account for deformations of the point-sets. Consequently, a general point feature registration algorithm needs to address all these issues. It should be able to solve for the correspondences between two point-sets, reje

13、ct out- liers and determine a good non-rigid transformation that can map one point-set onto the other. The need for non-rigid registration occurs in many real world applications. Tasks like template matching for hand-wri

14、tten characters in OCR, generating smoothly in- terpolated intermediate frames between the key frames in cartoon animation, track- ing human body motion in motion tracking, recovering dynamic motion of the heart in cardi

15、ac image analysis and registering human brain MRI images in brain map- ping, all involve finding the optimal transformation between closely related but dif- ferent objects or shapes. It is such a commonly occurring probl

16、em that many methods have been proposed to attack various aspects of the problem. However, be- cause of the great complexity introduced by the high dimensionality of the non-rigid mappings, all existing methods usually s

17、implify the problem to make it more tracta- ble. For example, the mappings can be approximated by articulated rigid mappings instead of being fully non-rigid. Restricting the point-sets to lie along curves, the set of co

18、rrespondence can be constrained using the curve ordering information. Simple heuristics such as using nearest-neighbor relationships to assign correspondence (as in the iterated closest point algorithm [5]) have also bee

19、n widely used. Though most methods tolerate a certain amount of noise, they normally assume that there are no outliers. These simplifications may alleviate the difficulty of the matching problem,H. Chui, A. Rangarajan /