Oral cbct motion artifact correction method, device, equipment, medium and product
By acquiring three-dimensional data and dental arch information from an oral CBCT system, and using neural networks for motion artifact correction, the complexity of marker introduction and high computational complexity in existing technologies are solved, achieving fast and accurate artifact correction results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TSINGHUA UNIVERSITY
- Filing Date
- 2025-11-03
- Publication Date
- 2026-07-21
AI Technical Summary
In the existing technology, motion artifact correction methods for dental CBCT systems require additional markers, which increases the complexity and cost of clinical operations. At the same time, image processing-based methods have high computational complexity and depend on the design and initial values of the objective function, making them difficult to implement in real time.
By acquiring three-dimensional oral scan data and three-dimensional surface data of the dental arch, initial oral motion parameters are determined. Neural networks are used for feature extraction and loss function optimization, avoiding the introduction of additional markers and directly using dental arch information for motion artifact correction.
It achieves fast and accurate motion artifact correction, avoids the use of additional markers, reduces computational complexity and cost, and improves correction efficiency.
Smart Images
Figure CN121242616B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, and in particular to a method, apparatus, equipment, medium, and product for correcting motion artifacts in oral CBCT. Background Technology
[0002] Cone-beam computed tomography (CBCT) is a crucial tool in modern dental diagnosis and treatment planning, widely used in implant planning, orthodontic analysis, periodontal disease diagnosis, and maxillofacial surgery due to its ability to provide high-quality three-dimensional maxillofacial anatomy. However, unavoidable physiological movements of the patient during CBCT scanning (such as swallowing, masticatory muscle contraction, and involuntary head movements) and the mechanical vibration of the equipment itself can lead to inconsistencies in the projection data acquisition process, resulting in motion artifacts in the reconstructed images. These motion artifacts manifest as stripes, double contours, or blurred anatomical structures, severely degrading image quality and affecting the accuracy of clinical diagnosis and the reliability of treatment planning. The formation mechanism of motion artifacts mainly stems from the geometric characteristics of CBCT scanning. During the CBCT scan, the X-ray source and detector rotate around the patient's head, acquiring hundreds of two-dimensional projection images, which are then used to generate three-dimensional volumetric data through reconstruction algorithms (such as the FDK algorithm). When the patient moves during the scan, the geometric consistency between the projection data is disrupted, leading to distortion in the reconstructed images.
[0003] Currently, in existing technologies, motion artifact correction in dental CBCT systems includes (1) a correction method based on markers and hardware assistance, which typically involves placing high-contrast reference markers (such as small metal balls) inside the patient's mouth or on the body surface. During the scan, by tracking the positional changes of these markers in a series of projected images, the rigid motion trajectory of the patient's head (usually containing three translational degrees of freedom and three rotational degrees of freedom) can be reconstructed. Subsequently, during image reconstruction, these motion trajectories are used to correct the projection geometry, thereby compensating for the effects of motion. (2) a correction method based on image processing and motion estimation, the core of which is the optimization problem of motion estimation. The basic process typically includes: initially reconstructing an image containing motion artifacts; generating simulated projection data through forward projection; comparing the simulated projection with the actual acquired projection data (usually constructing an objective function based on edge features, consistency conditions, etc.); adjusting motion parameters through iterative optimization algorithms to minimize the objective function, thereby finding the most probable motion trajectory; and finally using the estimated motion parameters for motion compensation reconstruction. Thirdly, there are deep learning-based correction methods. These methods avoid complex explicit motion modeling and instead allow the neural network to directly learn the features of motion artifacts and their removal methods from a large amount of data.
[0004] However, for marker- and hardware-assisted correction methods, additional marker placement steps are required, increasing the complexity of clinical procedures. Secondly, the positioning accuracy and stability of the markers themselves directly affect the correction effect. Furthermore, some optical tracking systems are expensive and difficult to integrate within the limited space of a dental clinic. For correction methods based on image processing and motion estimation, the computational complexity is high, the iteration process is time-consuming, and the optimization effect largely depends on the design of the objective function and the selection of initial values, posing challenges for real-time clinical applications. Summary of the Invention
[0005] This invention provides a method, apparatus, device, medium, and product for correcting motion artifacts in oral CBCT scans. It addresses the shortcomings of existing technologies where the introduction of additional markers leads to inaccurate correction and increased correction costs due to the introduction of hardware. The method acquires three-dimensional intraoral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity. It then determines initial oral motion parameters based on the three-dimensional intraoral scan data and three-dimensional surface data of the dental arch; determines forward projection data of oral motion based on the three-dimensional intraoral scan data and the initial oral motion parameters; and determines motion artifact correction parameters for the target oral cavity based on the forward projection data, CBCT projection data, and the initial oral motion parameters. Finally, it obtains the rigid motion trajectory of the target oral cavity based on the motion artifact correction parameters. This method avoids the introduction of additional markers and, because it directly incorporates accurate dental arch information, avoids the use of data containing motion artifacts, thus achieving a faster correction calculation speed without introducing new markers.
[0006] This invention provides a method for correcting motion artifacts in oral CBCT, comprising the following steps.
[0007] Acquire 3D oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity;
[0008] Initial oral motion parameters are determined based on three-dimensional intraoral scan data and three-dimensional surface data of the dental arch.
[0009] The anterior projection data of oral motion is determined based on the three-dimensional data of oral scanning and the initial oral motion parameters;
[0010] The motion artifact correction parameters for the target oral cavity are determined based on the anterior projection data of oral motion, CBCT projection data, and initial oral motion parameters.
[0011] According to the present invention, a method for correcting motion artifacts in oral CBCT is provided, which determines motion artifact correction parameters for the target oral cavity based on oral motion forward projection data, CBCT projection data, and initial oral motion parameters, including:
[0012] The loss function is determined based on oral motor projection data, CBCT projection data, and neural network.
[0013] The motion artifact correction parameters for the target oral cavity are determined based on the loss function and the initial oral motion parameters.
[0014] According to the present invention, a method for correcting motion artifacts in oral CBCT is provided, which determines a loss function based on oral motion forward projection data, CBCT projection data, and a neural network, including:
[0015] The oral cavity motion forward projection data is input into the neural network to obtain the forward projection coding features output by the neural network;
[0016] Input the CBCT projection data into the neural network to obtain the actual projection coding features output by the neural network;
[0017] The loss function is determined based on the forward projection coding features, the actual projection coding features, the true values of translation loss and rotation loss.
[0018] According to the present invention, a method for correcting motion artifacts in oral CBCT is provided, which determines a loss function based on forward projection coding features, actual projection coding features, translation loss ground truth, and rotation loss ground truth, including:
[0019] The inference training process loss is determined based on the forward projection coding features and the actual projection coding features.
[0020] The translation loss and rotation loss are determined based on the loss during the inference training process.
[0021] The loss function is determined based on the translation loss, rotation loss, the true value of translation loss, and the true value of rotation loss.
[0022] According to the present invention, a method for correcting motion artifacts in oral CBCT, after determining the motion artifact correction parameters of the target oral cavity based on a loss function and initial oral motion parameters, further includes:
[0023] Based on motion artifact correction parameters and intraoral 3D scan data, the forward projection data of oral motion is determined, and the step of determining the loss function based on the forward projection data of oral motion, CBCT projection data and neural network is continued.
[0024] According to the present invention, a method for correcting motion artifacts in oral CBCT is provided to determine three-dimensional surface data of the dental arch, including:
[0025] Acquire CBCT projection data of the target oral cavity;
[0026] Preliminary reconstruction of CBCT projection data is performed to obtain CBCT reconstructed data;
[0027] The CBCT reconstruction data were formally transformed to determine the three-dimensional surface data of the dental arch.
[0028] According to the present invention, a method for correcting motion artifacts in oral CBCT includes initial oral motion parameters, which include initial translation parameters and initial rotation parameters. The initial oral motion parameters are determined based on three-dimensional oral scan data and three-dimensional surface data of the dental arch, including:
[0029] Determine the point cloud registration algorithm;
[0030] The initial translation and rotation parameters are determined based on the three-dimensional surface data of the dental arch, the three-dimensional data of the intraoral scan, and the point cloud registration algorithm.
[0031] According to the present invention, a method for correcting motion artifacts in oral CBCT is provided, which acquires three-dimensional oral scan data of a target oral cavity, including:
[0032] Acquire intraoral scan data of the target oral cavity;
[0033] The surface scan data is processed to obtain the surface scan 3D data.
[0034] According to the present invention, a method for correcting motion artifacts in oral CBCT includes solid processing comprising repair processing and filling processing; solid processing is performed on oral scan data to obtain three-dimensional oral scan data, including:
[0035] The oral scan data is repaired to obtain candidate oral scan data;
[0036] The candidate oral scan data is filled to obtain the oral scan 3D data.
[0037] The present invention also provides a device for correcting motion artifacts in oral CBCT, comprising the following modules:
[0038] The 3D data acquisition module is used to acquire the 3D data of the target oral cavity and the cone-beam computed tomography (CBCT) projection data of the target oral cavity.
[0039] The initial parameter determination module is used to determine the initial oral motion parameters based on the three-dimensional data of the intraoral scan and the three-dimensional surface data of the dental arch.
[0040] The projection data determination module is used to determine the forward projection data of oral motion based on the three-dimensional data of oral scanning and the initial oral motion parameters.
[0041] The correction parameter determination module is used to determine the motion artifact correction parameters of the target oral cavity based on the oral cavity motion anterior projection data, CBCT projection data, and initial oral cavity motion parameters.
[0042] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described methods for correcting motion artifacts in oral CBCT.
[0043] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for correcting motion artifacts in oral CBCT.
[0044] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements any of the above-described methods for correcting motion artifacts in oral CBCT.
[0045] This invention provides a method, apparatus, device, medium, and product for correcting motion artifacts in oral CBCT. It involves acquiring three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity; determining initial oral motion parameters based on the three-dimensional oral scan data and three-dimensional surface data of the dental arch; determining forward projection data of oral motion based on the three-dimensional oral scan data and the initial oral motion parameters; and determining motion artifact correction parameters for the target oral cavity based on the forward projection data of oral motion, the CBCT projection data, and the initial oral motion parameters. The technical solution of this invention acquires three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity; determines initial oral motion parameters based on the three-dimensional oral scan data and three-dimensional surface data of the dental arch; determines forward projection data of oral motion based on the three-dimensional oral scan data and the initial oral motion parameters; determines motion artifact correction parameters of the target oral cavity based on the forward projection data of oral motion, CBCT projection data, and the initial oral motion parameters; and finally obtains the rigid motion trajectory of the target oral cavity based on the motion artifact correction parameters. This method avoids the introduction of additional markers, and because it directly incorporates accurate dental arch information, it can achieve a faster correction calculation speed without introducing new markers. Attached Figure Description
[0046] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0047] Figure 1 This is a flowchart illustrating the oral CBCT motion artifact correction method provided by the present invention.
[0048] Figure 2This is a schematic diagram of the structure of the oral CBCT motion artifact correction device provided by the present invention.
[0049] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation
[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0051] The following is combined with Figure 1 The present invention describes the oral CBCT motion artifact correction method provided by the present invention. The oral CBCT motion artifact correction method provided by the present invention is applicable to dental CBCT motion artifact correction based on the registration of oral scan three-dimensional data and dental arch three-dimensional surface data. The execution subject of this method can be an electronic device or an oral CBCT motion artifact correction device installed in the electronic device. The oral CBCT motion artifact correction device can be implemented by software, hardware or a combination of both.
[0052] Figure 1 This is a flowchart illustrating the oral CBCT motion artifact correction method provided by the present invention, as shown below. Figure 1 As shown, the method includes the following steps 101, 102, 103 and 104.
[0053] Step 101: Obtain the three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity.
[0054] In this step, the target oral cavity is the oral cavity that needs to be scanned. The three-dimensional oral scan data is the three-dimensional point cloud data of the target oral cavity obtained based on a high-precision oral scanner. The three-dimensional point cloud data can be, for example, the point cloud data of the patient's teeth in the target oral cavity. This embodiment does not limit this.
[0055] CBCT projection data can be, for example, a set of CBCT projection data of the target oral cavity that was actually acquired.
[0056] Specifically, the method involves acquiring three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity.
[0057] In one specific embodiment, acquiring three-dimensional intraoral scan data of the target oral cavity includes:
[0058] Acquire intraoral scan data of the target oral cavity;
[0059] The surface scan data is processed to obtain the surface scan 3D data.
[0060] In this step, the intraoral scan data is generally saved as a 3D model file format, such as .obj or .stl files. This format differs from the voxel format of the subsequently determined 3D surface data of the dental arch, representing different modalities. Therefore, the intraoral scan data needs to be preprocessed to obtain the same data format as the 3D surface data of the dental arch reconstructed by CBCT. Specifically, the intraoral scan data generally only contains surface information of the dental arch and needs to be converted from the 3D model file format to voxel format for projection. Furthermore, coarse registration is required before rigid registration to roughly align the intraoral scan data with the 3D surface data of the dental arch reconstructed by CBCT, providing an initial value for the rigid registration algorithm. This process uses point cloud registration, thus requiring the initial reconstructed image from CBCT to be converted into 3D surface data of the dental arch (CBCT preprocessing). Ultimately, the initial reconstructed images (voxel format of CBCT data), three-dimensional surface data of the dental arch (point cloud format of CBCT data), intraoral scan data (voxel format of intraoral scan data), and three-dimensional intraoral scan data (point cloud format of intraoral scan data) were obtained. These initial reconstructed images (voxel format of CBCT data), three-dimensional surface data of the dental arch (point cloud format of CBCT data), intraoral scan data (voxel format of intraoral scan data), and three-dimensional intraoral scan data (point cloud format of intraoral scan data) constituted the digital dental arch model. The digital dental arch model was used to determine the initial oral motion parameters in the subsequent process.
[0061] The rigid registration algorithm mainly uses a neural network to extract features, then determines the loss function by automatic differentiation, and determines the motion artifact correction parameters of the target oral cavity based on the loss function and the initial oral motion parameters. The neural network can be, for example, a neural network containing an encoder, etc., but this embodiment does not limit it.
[0062] Oral scan data is a three-dimensional surface model file, which is the initial three-dimensional point cloud data of the target oral cavity obtained by a high-precision oral scanner. The initial three-dimensional point cloud data can be, for example, the initial patient dental arch point cloud data of the target oral cavity, but this embodiment does not limit it.
[0063] Entity processing can employ various algorithms, such as triangular mesh volume generation algorithms, voxelization algorithms, tetrahedral mesh generation algorithms, implicit surface-based algorithms, etc. This embodiment does not limit these algorithms.
[0064] Specifically, a high-precision oral scanner is used to scan the complete dental arch of the target oral cavity, including the complete occlusal surfaces of the teeth, to obtain oral scan data of the target oral cavity. After obtaining the oral scan data of the target oral cavity, solid processing is performed on the oral scan data, that is, the oral scan data is filled into solids to obtain three-dimensional oral scan data.
[0065] In one specific embodiment, entity processing includes repair processing and filling processing; performing entity processing on oral scan data to obtain oral scan 3D data includes: performing repair processing on oral scan data to obtain candidate oral scan data; and performing filling processing on candidate oral scan data to obtain oral scan 3D data.
[0066] In this step, the repair process specifically involves model trimming and boundary repair of the intraoral scan data. This is achieved by using interactive selection tools or neural network-based automatic segmentation algorithms to identify and remove redundant non-dental arch components (such as soft tissue) from the intraoral scan data. Then, for holes and gaps on the surface of the 3D surface model containing the intraoral scan data, methods such as hole repair algorithms based on normal extension are used for repair. Ultimately, the goal is to ensure that the model is a complete and closed manifold surface, thus obtaining candidate intraoral scan data. This embodiment does not impose limitations on this process.
[0067] The filling process involves identifying the opening boundary of the dental arch model where the candidate oral scan data is located. First, the opening boundary loop is identified using algorithms such as minimum spanning tree or α-shape. This embodiment does not limit this process.
[0068] Subsequently, a reference plane is constructed for the opening boundary ring using plane fitting based on the least squares method, thereby forming a closed "bottom cover". This embodiment does not limit this.
[0069] After obtaining a completely closed surface, the space is filled using ray casting or flooding filling algorithms to obtain the three-dimensional data of the aperture scan. This embodiment does not limit this process.
[0070] Specifically, the oral scan data is repaired to obtain candidate oral scan data; the candidate oral scan data is then filled to obtain 3D oral scan data.
[0071] Step 102: Determine the initial oral motion parameters based on the three-dimensional surface data of the dental arch and the three-dimensional data of the oral scan.
[0072] In this step, the three-dimensional surface data of the dental arch is the data after format conversion of the reconstructed image with motion artifacts. The reconstructed image can be, for example, an image obtained based on the FDK algorithm, which stands for Feldkamp-Davis-Kress algorithm (an algorithm for three-dimensional image reconstruction). This embodiment does not limit this.
[0073] Specifically, after obtaining the 3D surface data of the dental arch and the STL data (i.e., 3D intraoral scan data) from the FDK-converted dental arch data and the 3D intraoral scan data, the 3D surface data of the dental arch and the 3D intraoral scan data are input into the point cloud registration module. The point cloud registration module can output 6 initial motion parameters, which are then used to determine the initial oral motion parameters.
[0074] For example, the six initial motion parameters could be: , They represent along Translation parameters of direction, Indicated by The rotation parameters are for the axis. After obtaining the three-dimensional surface data of the dental arch and the three-dimensional data of the oral scan, the oral scan data is used as the data to be registered, and the three-dimensional surface data of the dental arch is used as the fixed data. The translation and rotation parameters required for registering the data to be registered with the fixed data are obtained through the registration process, which are the 6 initial motion parameters.
[0075] The advantage of this setup is that the initial oral motion parameters can provide better initial values for determining motion artifact correction parameters based on the iterative rigid registration algorithm, thereby accelerating convergence.
[0076] In one specific embodiment, determining the three-dimensional surface data of the dental arch includes:
[0077] Acquire CBCT projection data of the target oral cavity;
[0078] Preliminary reconstruction of CBCT projection data is performed to obtain CBCT reconstructed data;
[0079] The CBCT reconstruction data were transformed to obtain three-dimensional surface data of the dental arch.
[0080] In this step, the form transformation is generally performed by extracting isosurfaces or calculating gradients. For example, the form transformation algorithm can be the marching cubes algorithm to extract isosurfaces, the Burson disk sampling algorithm to extract surface points, the gradient thresholding algorithm to extract surfaces, etc. This embodiment does not limit this.
[0081] Specifically, the process involves acquiring CBCT projection data of the target oral cavity; performing preliminary reconstruction on the CBCT projection data to obtain CBCT reconstruction data; and transforming the CBCT reconstruction data to determine the three-dimensional surface data of the dental arch.
[0082] In one specific embodiment, the initial oral motion parameters include initial translation parameters and initial rotation parameters; the initial oral motion parameters are determined based on intraoral 3D data and dentition 3D surface data, including:
[0083] Determine the point cloud registration algorithm;
[0084] The initial translation and rotation parameters are determined based on the three-dimensional surface data of the dental arch, the three-dimensional data of the intraoral scan, and the point cloud registration algorithm.
[0085] In this step, the point cloud registration module can use a variety of point cloud registration algorithms, including but not limited to the Iterative Closest Point (ICP) algorithm, a type of point cloud registration algorithm, coarse registration algorithm based on feature description, registration algorithm based on marker points, etc. This embodiment does not limit the specific algorithm used.
[0086] Specifically, the point cloud registration algorithm is determined; the three-dimensional surface data of the dental arch and the three-dimensional data of the intraoral scan are input into the point cloud registration algorithm for point cloud registration, thereby determining the initial oral motion parameters, which include initial translation parameters and initial rotation parameters.
[0087] Step 103: Determine the forward projection data of oral motion based on the three-dimensional data of oral scanning and the initial oral motion parameters.
[0088] Specifically, after obtaining the three-dimensional oral scan data and initial oral motion parameters, the three-dimensional oral scan data and initial oral motion parameters are subjected to differentiable forward projection to determine the forward projection data of oral motion.
[0089] Step 104: Determine the motion artifact correction parameters for the target oral cavity based on the oral cavity motion forward projection data, CBCT projection data, and initial oral cavity motion parameters.
[0090] Specifically, after obtaining the oral cavity motion forward projection data, CBCT projection data, and initial oral cavity motion parameters, feature extraction is performed on the oral cavity motion forward projection data and CBCT projection data respectively. Then, based on the results of feature extraction and the initial oral cavity motion parameters, the motion artifact correction parameters of the target oral cavity are determined.
[0091] There are various methods for feature extraction, such as using neural networks to extract features, but this embodiment does not limit the specific methods used.
[0092] In one specific implementation, motion artifact correction parameters for the target oral cavity are determined based on oral motion anterior projection data, CBCT projection data, and initial oral motion parameters, including:
[0093] The loss function is determined based on oral motor projection data, CBCT projection data, and neural network.
[0094] The motion artifact correction parameters for the target oral cavity are determined based on the loss function and the initial oral motion parameters.
[0095] In this step, the neural network can be, for example, a tool for extracting features from oral motion projection data and CBCT projection data, such as a neural network containing an encoder; this embodiment does not limit this.
[0096] Specifically, after obtaining the forward projection data of oral motor movements and CBCT projection data Next, the loss function was determined based on the oral motor projection data, CBCT projection data, and the neural network. Specifically, the oral motor projection data... and CBCT projection data Each parameter is input into its respective neural network to determine the loss function. The initial oral motion parameters are then updated based on the loss function. This process is iterated until convergence is achieved to obtain the motion artifact correction parameters for the target oral cavity.
[0097] In one specific embodiment, the loss function is determined based on oral motor projection data, CBCT projection data, and a neural network, including:
[0098] The oral cavity motion forward projection data is input into the neural network to obtain the forward projection coding features output by the neural network;
[0099] Input the CBCT projection data into the neural network to obtain the actual projection coding features output by the neural network;
[0100] The loss function is determined based on the forward projection coding features, the actual projection coding features, the true values of translation loss and rotation loss.
[0101] Specifically, oral motion forward projection data is input into a neural network, which then extracts features from the data to obtain the forward projection encoded features output by the neural network. The CBCT projection data is input into a neural network, which extracts features from the CBCT projection data to obtain the actual projection coding features output by the neural network. Then, the loss function is determined based on the forward projection coding features, the actual projection coding features, the true values of translation loss and rotation loss.
[0102] In one specific embodiment, the loss function is determined based on the forward projection coding features, the actual projection coding features, the translation loss truth value, and the rotation loss truth value, including:
[0103] The inference training process loss is determined based on the forward projection coding features and the actual projection coding features.
[0104] The translation loss and rotation loss are determined based on the loss during the inference training process.
[0105] The loss function is determined based on the translation loss, rotation loss, the true value of translation loss, and the true value of rotation loss.
[0106] In this step, the loss function is generally used to correct the network parameters of the neural network. To determine the loss function, the loss function... = ,in, Indicates the predicted loss. Represent the true value of loss and predict the loss. Includes predicted translation loss and predicted rotation loss, true value of loss. This includes the true values for translation loss and rotation loss. Then the loss function... Expanding the translation and rotation components, we obtain the loss function as follows: ,in, This represents the obtained loss function. Indicates translation loss. This represents the true value of the translation loss. Indicates rotational loss. Indicates the true value of rotational loss, subscript Subscript represents translation. This represents rotational loss.
[0107] Specifically, features are extracted from the forward projection data of oral motor movements using a neural network to obtain the forward projection encoded features output by the neural network. The CBCT projection data is input into a neural network, which extracts features from the CBCT projection data to obtain the actual projection coding features output by the neural network. Next, the inference training process loss is determined based on the forward projection encoding features and the actual projection encoding features. Specifically, the inference training process loss is used to correct for artifacts in the initial motion parameters. Then, the loss during the reasoning training process. Taking the derivative, we obtain the translation loss. The rotational loss is obtained. .in, Indicates the initial translation loss. This represents the initial rotational loss.
[0108] In one specific implementation, after obtaining the loss function, the initial oral motion parameters are updated using the loss function, and the above processing steps are iterated sequentially until final convergence, thereby determining the motion artifact correction parameters of the target oral cavity.
[0109] In one specific embodiment, after determining the motion artifact correction parameters for the target oral cavity based on the loss function and initial oral cavity motion parameters, the method further includes:
[0110] Based on motion artifact correction parameters and intraoral 3D scan data, the forward projection data of oral motion is determined, and the step of determining the loss function based on the forward projection data of oral motion, CBCT projection data and neural network is continued.
[0111] Specifically, after determining the motion artifact correction parameters of the target oral cavity based on the loss function and the initial oral motion parameters, iterative convergence can be further performed based on the motion artifact correction parameters. Based on the motion artifact correction parameters, the forward projection data of oral motion is determined with the 3D intraoral scan data, and the steps of determining the loss function based on the forward projection data of oral motion, CBCT projection data, and neural network are continued. The motion artifact correction parameters of the target oral cavity are determined based on the loss function and the initial oral motion parameters until the motion artifact correction parameters meet the preset parameter threshold. Alternatively, the number of iterations can be set. When the number of iterations reaches the required number, the correction parameters obtained corresponding to the current number of executions are determined as the motion artifact correction parameters. This embodiment does not limit this.
[0112] This invention provides a method for correcting motion artifacts in oral cavity using CBCT. The method involves acquiring three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity; determining initial oral motion parameters based on the three-dimensional oral scan data and three-dimensional surface data of the dental arch; determining forward projection data of oral motion based on the three-dimensional oral scan data and the initial oral motion parameters; and determining motion artifact correction parameters of the target oral cavity based on the forward projection data of oral motion, CBCT projection data, and the initial oral motion parameters. The technical solution of this invention utilizes three-dimensional surface data of the dental arch, three-dimensional oral scan data, and CBCT projection data of the target oral cavity; determines initial oral motion parameters based on the three-dimensional surface data of the dental arch and the three-dimensional oral scan data; determines forward projection data of oral motion based on the three-dimensional oral scan data and the initial oral motion parameters; and determines motion artifact correction parameters of the target oral cavity based on the forward projection data of oral motion, CBCT projection data, and the initial oral motion parameters. Finally, the rigid motion trajectory of the target oral cavity is obtained based on the motion artifact correction parameters. This method avoids introducing additional markers and, because it directly incorporates accurate dental arch information, achieves a relatively fast correction calculation speed without introducing new markers.
[0113] The oral CBCT motion artifact correction device provided by the present invention is described below. The oral CBCT motion artifact correction device described below can be referred to in correspondence with the oral CBCT motion artifact correction method described above.
[0114] Figure 2 This is a schematic diagram of the structure of the oral CBCT motion artifact correction device provided by the present invention, with reference to... Figure 2As shown, the oral CBCT motion artifact correction device 200 includes: a three-dimensional data acquisition module 201, an initial parameter determination module 202, a projection data determination module 203, and a correction parameter determination module 204; wherein,
[0115] The three-dimensional data acquisition module 201 is used to acquire the three-dimensional data of the target oral cavity and the cone-beam computed tomography (CBCT) projection data of the target oral cavity.
[0116] The initial parameter determination module 202 is used to determine the initial oral motion parameters based on the three-dimensional data of the oral scan and the three-dimensional surface data of the dental arch.
[0117] The projection data determination module 203 is used to determine the forward projection data of oral motion based on the three-dimensional data of oral scanning and the initial oral motion parameters.
[0118] The correction parameter determination module 204 is used to determine the motion artifact correction parameters of the target oral cavity based on the oral cavity motion forward projection data, CBCT projection data and initial oral cavity motion parameters.
[0119] In one example embodiment, the calibration parameter determination module 204 is specifically used for:
[0120] The loss function is determined based on oral motor projection data, CBCT projection data, and neural network.
[0121] The motion artifact correction parameters for the target oral cavity are determined based on the loss function and the initial oral motion parameters.
[0122] In one example embodiment, the correction parameter determination module 204 determines a loss function based on oral motion anterior projection data, CBCT projection data, and a neural network, specifically for:
[0123] The oral cavity motion forward projection data is input into the neural network to obtain the forward projection coding features output by the neural network;
[0124] Input the CBCT projection data into the neural network to obtain the actual projection coding features output by the neural network;
[0125] The loss function is determined based on the forward projection coding features, the actual projection coding features, the true values of translation loss and rotation loss.
[0126] In one example embodiment, the correction parameter determination module 204 determines a loss function based on the forward projection coding features, the actual projection coding features, the translation loss truth value, and the rotation loss truth value, specifically for:
[0127] The inference training process loss is determined based on the forward projection coding features and the actual projection coding features.
[0128] The translation loss and rotation loss are determined based on the loss during the inference training process.
[0129] The loss function is determined based on the translation loss, rotation loss, the true value of translation loss, and the true value of rotation loss.
[0130] In one example embodiment, the apparatus further includes a continuation execution module. The continuation execution module is configured to:
[0131] After determining the motion artifact correction parameters of the target oral cavity based on the loss function and the initial oral motion parameters, the forward projection data of oral motion is determined based on the motion artifact correction parameters and the three-dimensional intraoral scan data. Then, the step of determining the loss function based on the forward projection data of oral motion, CBCT projection data and neural network is continued.
[0132] In one example embodiment, the apparatus further includes a surface data determination module. The surface data determination module is configured to:
[0133] Acquire CBCT projection data of the target oral cavity;
[0134] Preliminary reconstruction of CBCT projection data is performed to obtain CBCT reconstructed data;
[0135] The CBCT reconstruction data were formally transformed to determine the three-dimensional surface data of the dental arch.
[0136] In one example embodiment, the initial oral cavity motion parameters include initial translation parameters and initial rotation parameters.
[0137] In one example embodiment, the initial parameter determination module 202 is specifically used for:
[0138] Determine the point cloud registration algorithm;
[0139] The initial translation and rotation parameters are determined based on the three-dimensional surface data of the dental arch, the three-dimensional data of the intraoral scan, and the point cloud registration algorithm.
[0140] In one example embodiment, the three-dimensional data acquisition module 201 acquires three-dimensional intraoral scan data of the target oral cavity, specifically for:
[0141] Acquire intraoral scan data of the target oral cavity;
[0142] The surface scan data is processed to obtain the surface scan 3D data.
[0143] In one example embodiment, entity processing includes repair processing and filling processing.
[0144] In one example embodiment, the 3D data acquisition module 201 performs solid processing on the intraorbital scan data to obtain intraorbital scan 3D data, specifically for:
[0145] The oral scan data is repaired to obtain candidate oral scan data;
[0146] The candidate oral scan data is filled to obtain the oral scan 3D data.
[0147] The apparatus of this embodiment can be used to execute the method of any embodiment in the side embodiment of the oral CBCT motion artifact correction method. Its specific implementation process and technical effects are similar to those in the side embodiment of the oral CBCT motion artifact correction method. For details, please refer to the detailed description in the side embodiment of the oral CBCT motion artifact correction method, which will not be repeated here.
[0148] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention, such as... Figure 3 As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340. The processor 310, communications interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute a CBCT motion artifact correction method for the oral cavity. This method includes: acquiring three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity; determining initial oral cavity motion parameters based on the three-dimensional oral scan data and three-dimensional surface data of the dental arch; determining forward projection data of oral cavity motion based on the three-dimensional oral scan data and the initial oral cavity motion parameters; and determining motion artifact correction parameters for the target oral cavity based on the forward projection data of oral cavity motion, the CBCT projection data, and the initial oral cavity motion parameters.
[0149] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0150] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the oral CBCT motion artifact correction method provided by the above methods. The method includes: acquiring three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity; determining initial oral motion parameters based on the three-dimensional oral scan data and three-dimensional surface data of the dental arch; determining motion artifact correction parameters of the target oral cavity based on the forward projection data of oral motion, the CBCT projection data, and the initial oral motion parameters; and determining the motion artifact correction parameters of the target oral cavity based on a loss function and the initial oral motion parameters.
[0151] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the oral CBCT motion artifact correction method provided by the above methods. The method includes: acquiring three-dimensional oral scan data and cone-beam computed tomography (CBCT) projection data of the target oral cavity; determining initial oral motion parameters based on the three-dimensional oral scan data and three-dimensional surface data of the dental arch; determining forward projection data of oral motion based on the three-dimensional oral scan data and the initial oral motion parameters; and determining motion artifact correction parameters of the target oral cavity based on the forward projection data of oral motion, the CBCT projection data, and the initial oral motion parameters.
[0152] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0153] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for correcting motion artifacts in oral CBCT, characterized in that, include: Acquire three-dimensional oral scan data of the target oral cavity and cone-beam computed tomography (CBCT) projection data of the target oral cavity; Initial oral motion parameters are determined based on the oral scan 3D data and the 3D surface data of the dental arch. The oral motion forward projection data is determined based on the oral scan 3D data and the initial oral motion parameters; The motion artifact correction parameters for the target oral cavity are determined based on the oral cavity motion forward projection data, the CBCT projection data, and the initial oral cavity motion parameters.
2. The method for correcting motion artifacts in oral CBCT according to claim 1, characterized in that, The step of determining the motion artifact correction parameters for the target oral cavity based on the oral cavity motion forward projection data, the CBCT projection data, and the initial oral cavity motion parameters includes: The loss function is determined based on the oral cavity motion projection data, the CBCT projection data, and the neural network; The motion artifact correction parameters for the target oral cavity are determined based on the loss function and the initial oral motion parameters.
3. The method for correcting motion artifacts in oral CBCT according to claim 2, characterized in that, The step of determining the loss function based on the oral cavity motion forward projection data, the CBCT projection data, and the neural network includes: The oral cavity motion forward projection data is input into the neural network to obtain the forward projection coding features output by the neural network; The CBCT projection data is input into the neural network to obtain the actual projection coding features output by the neural network; The loss function is determined based on the forward projection coding features, the actual projection coding features, the translation loss truth value, and the rotation loss truth value.
4. The method for correcting motion artifacts in oral CBCT according to claim 3, characterized in that, Determining the loss function based on the forward projection coding features, the actual projection coding features, the translation loss ground truth, and the rotation loss ground truth includes: The inference training process loss is determined based on the forward projection coding features and the actual projection coding features. The translation loss and rotation loss are determined based on the loss during the inference training process. The loss function is determined based on the translation loss, the rotation loss, the true value of the translation loss, and the true value of the rotation loss.
5. The method for correcting motion artifacts in oral CBCT according to claim 2, characterized in that, After determining the motion artifact correction parameters of the target oral cavity based on the loss function and the initial oral cavity motion parameters, the method further includes: Based on the motion artifact correction parameters and the oral scan 3D data, the forward projection data of oral cavity motion is determined, and the step of determining the loss function based on the forward projection data of oral cavity motion, the CBCT projection data, and the neural network is continued.
6. The method for correcting motion artifacts in oral CBCT according to claim 1, characterized in that, Determining the three-dimensional surface data of the dental arch includes: Acquire CBCT projection data of the target oral cavity; The CBCT projection data is initially reconstructed to obtain CBCT reconstructed data; The CBCT reconstruction data is formally transformed to determine the three-dimensional surface data of the dental arch.
7. The method for correcting motion artifacts in oral CBCT according to claim 1, characterized in that, The initial oral motion parameters include initial translation parameters and initial rotation parameters; determining the initial oral motion parameters based on the oral scan 3D data and dentition 3D surface data includes: Determine the point cloud registration algorithm; The initial translation parameters and the initial rotation parameters are determined based on the three-dimensional surface data of the dental arch, the three-dimensional data of the intraoral scan, and the point cloud registration algorithm.
8. The method for correcting motion artifacts in oral CBCT according to claim 1, characterized in that, The acquisition of three-dimensional intraoral scan data of the target oral cavity includes: Acquire intraoral scan data of the target oral cavity; The oral scan data is processed to obtain the oral scan 3D data.
9. The method for correcting motion artifacts in oral CBCT according to claim 8, characterized in that, The entity processing includes repair processing and filling processing; the entity processing of the oral scan data to obtain the oral scan 3D data includes: The oral scan data is subjected to the repair process to obtain candidate oral scan data; The candidate oral scan data is subjected to the filling process to obtain the oral scan three-dimensional data.
10. A device for correcting motion artifacts in oral CBCT, characterized in that, include: The three-dimensional data acquisition module is used to acquire the three-dimensional data of the target oral cavity and the cone-beam computed tomography (CBCT) projection data of the target oral cavity; An initial parameter determination module is used to determine initial oral motion parameters based on the oral scan three-dimensional data and the three-dimensional surface data of the dental arch. The projection data determination module is used to determine the forward projection data of oral motion based on the oral scan three-dimensional data and the initial oral motion parameters; The correction parameter determination module is used to determine the motion artifact correction parameters of the target oral cavity based on the oral cavity motion forward projection data, the CBCT projection data, and the initial oral cavity motion parameters.
11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the oral CBCT motion artifact correction method as described in any one of claims 1 to 9.
12. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the oral CBCT motion artifact correction method as described in any one of claims 1 to 9.
13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the oral CBCT motion artifact correction method as described in any one of claims 1 to 9.