High-efficiency registration method aimed at CT and optical scanning tooth model

Abstract

The invention provides a high-efficiency registration method aimed at a CT and optical scanning tooth model, and belongs to the field of computer oral cavity recovery. The method comprises the steps of (1) data preparation in which 3D visualization is carried out on CT data of the oral cavity to obtain a CT reconstructed tooth image, visualization is carried out on optical scanning data of the tooth crown and gum part to obtain an optical scanning tooth crown image, and the CT reconstructed tooth image and the optical scanning tooth crown image are arranged in the unified world coordinate system; (2) initial registration in which characteristic points are selected manually to process the CT reconstructed tooth image and the optical scanning tooth crown image and further to obtain an initial registration model; and (3) accurate registration in which characteristic points are selected automatically to process the initial registration model and further to obtain an accurate registration model. Problems in registration data with scale transformation and different resolution are solved, and a registration result is high in convenience speed and low in registration error.

Description

technical field [0001] The invention belongs to the field of computer oral restoration, and in particular relates to a high-efficiency registration method for CT and optical scanning tooth models, which is used to form a three-dimensional tooth model with complete data and high precision. Background technique [0002] In the computerized oral implant restoration system, in the aided design of the invisible appliance without brackets, the most important thing is to obtain complete and high-precision three-dimensional digital data composed of CT root data and optical scanning crown data. Tooth model, because the accuracy and integrity of the model are directly related to the results of subsequent tooth arrangement, implantation, orthodontics, and biomechanical analysis (see "Wu T, Zhang L B. Three dimention tooth reconstruction using level set active contour model [J]. Journal of Image and Graphics, 2016, 21(8):1078-1087"). The three-dimensional imaging technologies commonly ...

AI Technical Summary

Problems solved by technology

During the registration process, two main difficult problems arise: one is the selection of feature points, and the other is the merging of images of different modalities with different resolutions (refer to "Mei X, Li Z, Xu S, et al .Registration of the Cone Beam CT and blue-ray scanned dental model based on the improved ICP Algorithm[J].International Journal of Biomedical Imaging,2014:348740-348740”

However, in the process of extracting feature points with curvature features, the threshold must be properly adjusted according to the complexity and accuracy of the model, thus causing inconvenience to the operator

In 2015 and 2016, an improved ICP point cloud registration algorithm was proposed. This algorithm first uses the principal component analysis method to realize the initial registration of the point cloud, and obtains a better initial position of the point cloud. Then, on the basis of the ICP algorithm, it uses K-D The tree structure accelerates the search, and uses the direction vector angle threshold to remove wrong point pairs to improve the efficiency of the algorithm. Small, the algorithm failed to significantly improve the efficiency of accurate registration

In 2015, W et al. manually selected feature points from the two images, and then used the ICP algorithm for registration. However, the accuracy of the position mainly depends on the operator's experience and operating status, and the human factor is large, so it leads to registration. accurate inaccurate

[0006] 1) It is only for rigid body transformation, without introducing scale transformation parameters, and cannot solve the registration of two similar objects with different sizes and resolutions;

[0007] 2) The selection of feature points, when manually selecting feature points, there are large human interference factors; when automatically selecting feature points, the algorithm is more complicated and the time utilization rate is not high;

[0008] 3) In the search of the latest iterative point pair, all the data points of one piece of point cloud are searched for the nearest point in another piece of data, which is prone to errors in the nearest point pair, which has a great impact on the error of the registration result, and the amount of calculation is also large , low time efficiency

Images