Analysis Of Occlusal Contacts Based On Point Cloud Spatial Features

Automatic analysis of the occlusion relationship can reduce the labor time cost,improve the accuracy and precision of the medical effectiveness assessment,and meet the clinical needs of child orthodontics.The key step of the occlusal contacts analysis is the registration of the point cloud,which uses the overlapping part of the point cloud before and after treatment to find the correct correspondences.However,the point cloud before and after treatment dramatically change due to the noise generated during treatment,development and sampling.The noise may make the registration of the jaw point cloud very challenging.To deal with the challenges of the low-overlapping jaw point cloud registration,we apply the theory of WALA ridge to identify the overlapping region in the point cloud registration.The contributions of this work contains three parts:overlapping region localization,point cloud registration and occlusal contacts analysis.Overlapping region localization: This paper proposes to extract the skeleton line as the boundary line and use the normal distance as the criterion to obtain the overlapping region on one side of the skeleton line,overlapping region localization works well.This paper proposes the skeleton line extraction technology based on the projection matching group.The matching group contains the correspondences between point cloud and projection points,using the correspondences to map the two-dimensional skeleton line to the three-dimensional space.Through different data experiments,the skeleton line extracted by this method is more in line with the overall contour of the point cloud,and is robust to the interference of the edge shape.The effect is more concentrated in the center of the point cloud than the traditional scheme;In the case of different sampling,the time consumed by the algorithm in this paper is maintained at the same level.Point cloud registration: To improve the iterative nearest point algorithm,based on the convex feature of the WALA ridge,the convex density extreme is used as the constraint condition in the correspondences estimation,and the corresponding point pairs that do not meet the constraint condition are eliminated,so that the correspondence relationship of the overlapping regions is more accurate.Finally,the iterative optimal transformation is used to align the jaw point cloud.Through different data experiments,the registration accuracy is improved.For different overlap rate thresholds(?=0.1,?=0.5,?=1),the overlap rate exceeds the traditional algorithm;becausethe correspondences estimation are constrained,the number of corresponding point pairs is reduced,which reduces time spent by the algorithm.Occlusal contacts analysis: This paper designs and implements system for the occlusal contacts analysis.The occlusal surface is defined by the least squares method and the projection analysis is performed,the proposed approach realizes the automatic analysis of the occlusal contacts.Experiments show that the proposed approach also improves the efficiency,the accuracy and the precision of medical effectiveness assessment.It is believed that the proposed approach has promising application in basic research and diagnosis of the child jaw deformity.