WO2023165505A1 - 牙齿矫正效果的检测方法、装置、设备和存储介质 - Google Patents

牙齿矫正效果的检测方法、装置、设备和存储介质 Download PDF

Info

Publication number
WO2023165505A1
WO2023165505A1 PCT/CN2023/078959 CN2023078959W WO2023165505A1 WO 2023165505 A1 WO2023165505 A1 WO 2023165505A1 CN 2023078959 W CN2023078959 W CN 2023078959W WO 2023165505 A1 WO2023165505 A1 WO 2023165505A1
Authority
WO
WIPO (PCT)
Prior art keywords
tooth
target
image
model
information
Prior art date
Application number
PCT/CN2023/078959
Other languages
English (en)
French (fr)
Inventor
华昀峰
田彦
江腾飞
Original Assignee
先临三维科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 先临三维科技股份有限公司 filed Critical 先临三维科技股份有限公司
Publication of WO2023165505A1 publication Critical patent/WO2023165505A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/10—Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2200/00—Indexing scheme for image data processing or generation, in general
    • G06T2200/08—Indexing scheme for image data processing or generation, in general involving all processing steps from image acquisition to 3D model generation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G06T2207/10012—Stereo images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30036—Dental; Teeth
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2210/00—Indexing scheme for image generation or computer graphics
    • G06T2210/41—Medical

Definitions

  • the present disclosure relates to the technical field of information processing, and in particular to a detection method, device, equipment and storage medium for the effect of orthodontics.
  • 3D digital orthodontics as a new digital orthodontic technology, is widely used.
  • the doctor In the process of oral dental treatment, the doctor generally collects the user's tooth image through the handle of the intraoral scanner, then checks the tooth image displayed on the monitor, and judges whether the user's tooth is missing or defective after diagnosis.
  • the technical problem to be solved in the present disclosure is to solve the problem that the existing orthodontic treatment process is relatively time-consuming and the review cost is high.
  • an embodiment of the present disclosure provides a method for detecting the effect of orthodontics, including:
  • the tooth image includes the target tooth
  • the tooth model is mapped to the tooth image based on the pose information, and a detection result of the target tooth correction effect is generated based on the tooth image.
  • a device for detecting the effect of orthodontics including:
  • an acquisition unit configured to acquire a tooth image, the tooth image including target teeth
  • An identification unit configured to identify the target tooth in the tooth image based on the pre-trained neural network model, and determine the characteristic information of the target tooth
  • a determination unit configured to determine the pose information of the target tooth according to the feature information and the tooth model corresponding to the target tooth
  • a detection unit configured to map the tooth model to the tooth image based on the pose information, and generate a detection result of the target tooth correction effect based on the tooth image.
  • an electronic device including:
  • the computer program is stored in the memory and is configured to be executed by the processor to implement the above-mentioned method for detecting the effect of orthodontics.
  • a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned method for detecting the effect of orthodontics are implemented.
  • the method for detecting the orthodontic effect includes: acquiring a tooth image; identifying the tooth image based on a pre-trained neural network model, and determining the feature information of the target tooth in the tooth image; corresponding to the target tooth according to the feature information
  • the tooth model of the target tooth is determined to determine the pose information of the target tooth; the tooth model is mapped to the tooth image based on the pose information, and the detection result of the target tooth correction effect is generated based on the tooth image.
  • the relevant data of the orthodontic effect can be automatically obtained in real time only by collecting the image of the currently corrected tooth, and the test of the effect of the orthodontic is easy to operate and easy to implement, and it is also convenient for follow-up Relevant medical staff understand the correction process and adjust the correction plan in time.
  • FIG. 1 is a schematic flowchart of a method for detecting the effect of orthodontics provided by an embodiment of the present disclosure
  • FIG. 2 is a schematic diagram of an application scenario provided by an embodiment of the present disclosure
  • Fig. 3 is a schematic diagram of a tooth image provided by an embodiment of the present disclosure.
  • Fig. 4 is a schematic diagram of a tooth model provided by an embodiment of the present disclosure.
  • Fig. 5 is a schematic flowchart of a method for detecting the effect of orthodontics provided by an embodiment of the present disclosure
  • Fig. 6 is the structure of a device for detecting the effect of orthodontics provided by an embodiment of the present disclosure schematic diagram
  • FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • Orthodontic treatment mainly applies orthodontic force to the teeth to guide the remodeling of the periodontal tissue, thereby changing the position of the teeth in the alveolar bone.
  • Orthodontic treatment can improve the bad occlusal relationship caused by crowded dentition and abnormal arrangement of teeth, so as to achieve long-term stability of periodontal tissue.
  • 3D digital orthodontics is a new digital orthodontic technology, which is widely used. It is more accurate than traditional orthodontics in the correction of misaligned teeth and buck teeth. Its effect is more natural, the correction time is shorter, and it does not bounce back. Favored by those with beautiful teeth.
  • the doctor In the process of oral dental treatment, the doctor generally collects the user's tooth image through the handle of the intraoral scanner, then checks the tooth image displayed on the monitor, and judges whether the user's tooth is missing or defective after diagnosis.
  • the user In the process of traditional orthodontic treatment, the user needs to come to the hospital for follow-up visits many times. During each review, the user's teeth need to be scanned by an intraoral scanner to obtain and process the user's current tooth model, and then compare it with the preset model. Check whether the orthodontic condition matches the pre-designed plan to obtain error reports and treatment effects.
  • this process is time-consuming and labor-intensive, and the cost of re-examination is high, and the user's cooperation is not high. Delayed follow-up visits often occur , which leads to an increase in the failure rate of the correction plan, affects the final treatment effect, and is also a waste of manpower and material resources.
  • the embodiment of the present disclosure provides a method for detecting the effect of orthodontics.
  • the current orthodontic effect of the teeth is analyzed based on the tooth images and the expected model by taking real-time images of the teeth, which is easy to operate.
  • it can also provide correction data for relevant medical staff, so that subsequent adjustments can be made based on the correction data, and Frequent review is not required, which reduces the waste of manpower and material resources to a certain extent.
  • it will be described in detail through one or more of the following embodiments.
  • Figure 1 is a method for detecting the effect of orthodontics provided by an embodiment of the present disclosure, which is used to detect the current orthodontic effect of teeth, so as to facilitate the understanding of the orthodontic situation, specifically including the following steps S110 to S140 as shown in Figure 1:
  • the method for detecting the effect of orthodontics can be executed by a terminal or a server.
  • the terminal or server can detect the correction effect of the target tooth through the tooth model and the tooth image.
  • the target tooth refers to a single tooth, and a tooth image may include multiple teeth. When detecting each tooth, it can be called This tooth is the target tooth.
  • the server 22 includes a tooth model, and the server 22 receives the tooth image transmitted by the terminal 21, and then detects the orthodontic effect based on the tooth image and the tooth model.
  • the tooth image may be obtained by shooting by the terminal 21 .
  • the tooth image is obtained by the terminal 21 from other devices.
  • the tooth image is an image obtained by the terminal 21 performing image processing on a preset image
  • the preset image may be obtained by the terminal 21, or the preset image may be obtained by the terminal 21 from other devices.
  • other devices are not specifically limited.
  • the terminal 21 includes a tooth model, and after acquiring the tooth image, the terminal 21 detects the orthodontic effect according to the tooth image and the tooth model.
  • the method for detecting the orthodontic effect provided by the embodiments of the present disclosure is not limited to the several possible scenarios described above.
  • the method for detecting the orthodontic effect independently performed by the terminal 21 will be described below as an example.
  • the terminal obtains the tooth image including the target tooth through the acquisition device.
  • the tooth image includes at least one relatively complete tooth. If it includes multiple teeth, it is necessary to analyze the correction of each tooth in the multiple teeth.
  • a tooth image is shown in FIG. 3 .
  • FIG. 3 is a schematic diagram of a tooth image provided by an embodiment of the present disclosure.
  • the tooth image 310 in FIG. and tooth 312; tooth image 320 in FIG. 3 includes tooth 321.
  • the acquisition device An external device that can be embedded in the terminal or used in combination with the terminal.
  • the terminal can be a mobile phone or a tablet computer, and the device embedded in the terminal can be a device attached to the terminal itself, such as a camera configured in the terminal.
  • the camera captures two-dimensional images of teeth, Or take video data of a tooth.
  • Each frame of the video data is a two-dimensional tooth image.
  • the image of the frame, the specific external device is not limited. It is understandable that there is no limitation on the manner of acquiring the tooth image, and it can be selected according to the needs of the user.
  • the terminal needs to obtain a tooth model before obtaining the tooth image.
  • the method specifically includes: obtaining tooth scan data, and constructing a three-dimensional tooth model according to the tooth scan data.
  • the three-dimensional tooth model includes a tooth model corresponding to at least one tooth; the tooth scan Data were acquired by a 3D scanner.
  • the tooth model corresponding to each tooth in the three-dimensional tooth model is numbered.
  • the terminal will scan the teeth with a 3D scanning device to obtain the tooth scan data before taking the tooth image. Scanning, generating tooth scanning data, constructing the tooth model of the user's teeth before orthodontics according to the tooth scanning data, all single tooth models form the user's 3D tooth model, and can also adjust the pose of each tooth model before orthodontics to obtain
  • the three-dimensional tooth model after correction, all the tooth models in the three-dimensional tooth model after correction constitute the expected model, and the expected model is the most ideal effect model after correction.
  • the following embodiments take the tooth model after correction as an example for illustration.
  • the neural network model After obtaining the corrected tooth model of each tooth, because the whole 3D tooth model is obtained through scanning data, it is necessary to use the neural network model to divide the user's 3D tooth model (3D tooth model) into a single tooth model, that is, It is said that the three-dimensional tooth model is composed of multiple single tooth models, and each tooth model is numbered.
  • the number is convenient for positioning a single tooth model. For example, this user includes 28 teeth, and each tooth needs to be corrected. Number the 28 single tooth models in the tooth model, which can be recorded as 1 to 28 respectively. Understandably, it is also possible to number only the single tooth model that needs to be corrected, and not to number the rest of the teeth that do not need to be corrected.
  • FIG. 4 is a schematic diagram of a tooth model provided by an embodiment of the present disclosure.
  • FIG. 4 is a three-dimensional tooth model, and the three-dimensional tooth model is a three-dimensional model.
  • Teeth 311 and teeth 312 in tooth 300 in FIG. The tooth model is 413.
  • the feature information includes feature points, coordinates of the target tooth, and a serial number of the target tooth.
  • the terminal after the terminal acquires the tooth image, it recognizes the tooth image based on the pre-trained neural network model, and determines the feature information of the target tooth in the tooth image.
  • the pre-trained neural network model may be The model of the extraction and recognition function, the specific model is not limited, as long as the teeth can be identified to obtain the feature information; wherein, the feature information includes the feature points, numbers and coordinates of the tooth image, if the tooth image includes multiple teeth, then identify The feature information of each tooth among multiple teeth can also be obtained, and a feature information including all tooth features can also be obtained.
  • a feature information can also be understood as the feature information corresponding to the tooth image.
  • the tooth image in Figure 3 310 includes a tooth 311 and a tooth 312, and the tooth image 320 includes a tooth 321. If a single tooth is detected, both the tooth 311 and the tooth 312 in the tooth image 310 can be recorded as the target tooth, and the tooth 321 in the tooth image 320 is recorded as the target tooth.
  • the following embodiments take tooth 321 as an example for illustration.
  • the feature information corresponding to tooth 321 includes feature points of tooth 321. Feature points can be points on the tooth image where the gray value of the image changes sharply or on the edge of the tooth image.
  • the coordinates of the tooth 321 can specifically refer to a plurality of coordinate points included in the outline (edge line) of the tooth 321, and the number of the tooth 321 is used to determine a specific single tooth model in the pre-built three-dimensional tooth model, That is to say, the number of the tooth 321 is the same as that of a single tooth model in the above three-dimensional tooth model, which is convenient for determining the tooth model corresponding to the target tooth.
  • the tooth model corresponding to the tooth 321 is 412 in FIG. 4 .
  • the tooth model corresponding to the target tooth is determined according to the number in the feature information corresponding to the target tooth, and then the target tooth is determined according to the feature points and/or coordinates in the feature information and related information of the tooth model.
  • the pose information of the tooth includes orientation and spatial position information
  • the orientation refers to the posture of the tooth, for example, whether the tooth is facing up or down relative to the three-dimensional tooth model
  • the spatial position information can also be understood as three-dimensional coordinate information.
  • the method further includes: acquiring other tooth images including the target tooth, and according to the first feature point in the feature information corresponding to the other tooth image and the first feature point in the feature information corresponding to the tooth image The second feature point updates the pose information.
  • the pose information of the target tooth is directly calculated according to the coordinates in the feature information and the relevant information of the tooth model corresponding to the target tooth; if multiple images including the target tooth are obtained image, the pose information is calculated based on the coordinates corresponding to any one of the multiple dental images and the related information of the tooth model, and the pose information is updated based on the feature points corresponding to the multiple dental images, that is, by matching
  • the feature points detected in different tooth images optimize the determined pose information to improve accuracy.
  • the feature points corresponding to multiple tooth images include the first feature point included in any of the above tooth images, and multiple tooth images
  • the 10 pose information corresponding to the tooth 321 is calculated respectively, and then according to the feature points in the feature information corresponding to the 10 tooth images, combined
  • the 10 pose information determined above are optimized to obtain the optimal pose information, so as to improve the accuracy of the detection result.
  • the tooth image after projection includes both the two-dimensional tooth model and the target tooth, the image is a two-dimensional image, the tooth model is a three-dimensional tooth model, and the three-dimensional tooth model is mapped to a two-dimensional After the tooth image, the tooth image includes a two-dimensional tooth model.
  • the tooth model is projected into the tooth image 320 to obtain a mapping image 330.
  • the mapping image 330 includes a two-dimensional tooth model 331 and a target tooth 321.
  • the two-dimensional tooth model 331 maps the target tooth 321 most of the coverage. It is also possible to calculate the pixel difference between the tooth image and the mapping image, and use the difference between the target tooth in the tooth image and the tooth model pixels in the mapping image to reflect the correction of the teeth during the correction process.
  • the detection result can be whether the correction effect is correct or not.
  • the relevant data and test results involved in the test process can be sent to the relevant medical staff for inspection, and the relevant medical staff can judge whether the test results meet expectations, which is convenient for the follow-up to quickly understand the correction situation and adjust the correction method in time.
  • the tooth model is mapped to the tooth image based on the pose information, and the detection result of the target tooth correction effect is generated based on the tooth image, which specifically includes: mapping the tooth model to the tooth image based on the pose information to obtain the mapped image ; Calculate the pixel difference between the tooth image and the mapping image; generate the detection result of the target tooth correction effect according to the pixel difference.
  • the generation of detection results in the above S140 specifically includes the following implementation process: based on the pose information of the tooth image, the tooth model corresponding to the target tooth is mapped to the tooth image to obtain the mapped image, see Figure 3, the tooth image in Figure 3 320 includes tooth 321, after obtaining the pose information of tooth 321, the tooth model corresponding to tooth 321 is made according to the pose information Type 413 is mapped to the tooth image 320, and the tooth image after mapping, that is, the mapping image is denoted as 330, and the mapping image 330 includes the mapped two-dimensional tooth model 331 and the target tooth 321 covering most of the area by the tooth model 331; the mapping is obtained After the image 330, calculate the pixel difference between the target tooth 321 and the tooth model 331 in the mapping image 330, or directly calculate the pixel difference between the tooth image 320 and the mapping image 330, that is, calculate the pseudo tooth 321 of the target tooth 321 and the tooth model 331 Combined error to get the pixel difference, and finally analyze the
  • Whether it meets expectations refers to whether the current orthodontic situation of the teeth conforms to the corrected tooth model constructed above, that is, whether the current orthodontic situation is in line with The tooth model is the standard, and there is still a correction error, and then the detection result of the target tooth correction effect is generated according to the pixel difference and the analysis result.
  • a method for detecting the effect of orthodontics includes: acquiring a tooth image, which includes a target tooth; identifying the target tooth in the tooth image based on a pre-trained neural network model, and determining the characteristic information of the target tooth ; According to the feature information and the tooth model corresponding to the target tooth, determine the pose information of the target tooth; map the tooth model to the tooth image based on the pose information, and generate the detection result of the target tooth correction effect based on the tooth image.
  • the user in the process of dental diagnosis and treatment, the user can collect dental images by himself according to the mobile terminal.
  • the mobile terminal can be a mobile phone, and the user can complete the detection at home without using a professional intraoral scanner, so the user does not need to go frequently.
  • Medical obedience is also convenient for relevant medical staff to adjust the correction plan in time to improve the success rate of orthodontic correction.
  • the three-dimensional Select the tooth model with the same tooth number as the target tooth in the tooth model includes all single tooth models of the user's teeth to be corrected, and the single tooth models have numbers.
  • the target tooth 321 in the tooth image 320 in FIG. 3 The number is 16, and a single tooth model with the same number 16 is obtained in the three-dimensional tooth model, that is, the tooth model 413 in the above-mentioned FIG. 4 .
  • the pose information of the target tooth is determined according to the coordinates in the feature information and the related information of the tooth model.
  • determine the pose information of the target tooth according to the feature information and the tooth model specifically including: determining the coordinate system of the tooth model, and transforming the coordinates of the target tooth into the coordinate system; based on the transformed coordinates of the target tooth and The pose information of the tooth model determines the pose information of the target tooth in the coordinate system.
  • determining the pose information of the target tooth according to the feature information and the tooth model specifically includes the following steps: determining the coordinate system of the tooth model, the coordinate system is specifically the world coordinate system, and the coordinate system is determined by a three-dimensional scanner that scans the tooth After determining the coordinate system, transform the coordinates of the target tooth into the world coordinate system, the coordinates corresponding to the tooth image are the image coordinate system, and the coordinates of the target tooth in the tooth image are also the image coordinate system, and convert the coordinates of the target tooth to In the world coordinate system, the conversion between the image coordinate system and the world coordinate system can also be understood as unifying the entire tooth image to the world coordinate system, which is convenient for determining the pose information of the target tooth in the world coordinate system; then based on the conversion After the coordinates of the target tooth and the pose information of the tooth model in the world coordinate system, the pose information of the target tooth in the world coordinate system is determined.
  • the coordinates of the target tooth can specifically refer to the outline or edge of the target tooth in the tooth image.
  • determine the pose information of the target tooth in the coordinate system specifically including: based on the transformed coordinates of the target tooth and the position information of the tooth model, Fit the contour line of the target tooth with the contour line of the tooth model to determine the position information of the target tooth; according to the orientation of the tooth model, Determine the orientation of the target tooth; obtain the pose information of the target tooth in the coordinate system according to the orientation and position information of the target tooth.
  • determining the pose information of the target tooth in the coordinate system specifically includes the following steps: the pose information includes position information and orientation, and the position information is The three-dimensional coordinates in space, the orientation is used to determine the posture of the tooth in the entire tooth model; based on the converted two-dimensional coordinates of the target tooth contour line and the position information of the tooth model in the world coordinate system, the contour line of the target tooth and the tooth model Fit the contour line of the target tooth to determine the position information of the target tooth in this coordinate system; then determine the orientation of the target tooth in this coordinate system according to the orientation in the pose information of the tooth model; finally, according to the orientation of the target tooth and the target The position information of the tooth, the pose information of the target tooth in the coordinate system is obtained, and the conversion relationship between the tooth image and the tooth model is established after determining the pose information of the target tooth, which is convenient for subsequent mapping of the tooth model to the tooth image , to analyze the correction effect.
  • An embodiment of the present disclosure provides a method for detecting the effect of orthodontics.
  • Fig. 6 is a schematic structural diagram of a device for detecting a tooth correction effect provided by an embodiment of the present disclosure.
  • the device for detecting the effect of orthodontics provided by the embodiments of the present disclosure can execute the processing flow provided in the above embodiment of the method for detecting the effect of orthodontics.
  • the device for detecting the effect of orthodontics 600 includes:
  • An acquisition unit 610 configured to acquire a tooth image, where the tooth image includes a target tooth
  • the identification unit 620 is configured to identify the target tooth in the tooth image based on the pre-trained neural network model, and determine the characteristic information of the target tooth;
  • a determining unit 630 configured to use the feature information and the tooth corresponding to the target tooth A model for determining the pose information of the target tooth;
  • the detection unit 640 is configured to map the tooth model to the tooth image based on the pose information, and generate a detection result of the target tooth correction effect based on the tooth image.
  • the device 600 also includes a construction unit, which is used before the tooth image is acquired, specifically for:
  • the tooth scan data is acquired by a three-dimensional scanner.
  • the characteristic information in the device 600 includes the number of the target tooth; the tooth model corresponding to each tooth in the three-dimensional tooth model has a number.
  • the determination unit 630 determines the pose information of the target tooth according to the feature information and the tooth model corresponding to the target tooth, specifically for:
  • the feature information and the tooth model determine the pose information of the target tooth.
  • the feature information in the device 600 also includes the coordinates of the target teeth.
  • the determination unit 630 determines the pose information of the target tooth according to the feature information and the tooth model, specifically for:
  • the pose information in the apparatus 600 includes orientation and position information.
  • the determination unit 630 determines the pose information of the target tooth in the coordinate system based on the converted coordinates of the target tooth and the pose information of the tooth model, specifically for:
  • the pose information of the target tooth in the coordinate system is obtained according to the orientation and position information of the target tooth.
  • the feature information in device 600 further includes feature points, where the feature points are used to identify local features of the target tooth.
  • the device 600 also includes an update unit, which is used for determining the pose information of the target tooth after the update unit is specifically used for:
  • the detection unit 640 maps the tooth model to the tooth image based on the pose information, and generates the detection result of the target tooth correction effect based on the tooth image, specifically for:
  • a detection result of the target tooth correction effect is generated according to the pixel difference.
  • the device for detecting the orthodontic effect of the embodiment shown in FIG. 6 can be used to implement the technical solution of the above-mentioned method embodiment, and its implementation principle and technical effect are similar, and will not be repeated here.
  • FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • the electronic device provided by the embodiments of the present disclosure can execute the processing flow of the method for detecting the orthodontic effect provided by the above embodiments.
  • the electronic device 700 includes: a processor 710, a communication interface 720, and a memory 730; stored in the memory 730 and configured to be executed by the processor 710 as the method for detecting the effect of orthodontics described above.
  • an embodiment of the present disclosure also provides a computer-readable storage medium on which There is a computer program, and the computer program is executed by the processor to implement the method for detecting the orthodontic effect described in the above-mentioned embodiments.
  • an embodiment of the present disclosure also provides a computer program product, the computer program product includes a computer program or an instruction, and when the computer program or instruction is executed by a processor, the method for detecting an orthodontic effect as described above is implemented.
  • an embodiment of the present disclosure also provides a device, which includes:
  • memory for storing processor-executable instructions
  • the processor is configured as:
  • the tooth image includes the target tooth
  • the tooth model is mapped to the tooth image based on the pose information, and a detection result of the target tooth correction effect is generated based on the tooth image.
  • the detection method of the orthodontic effect provided by the disclosure can effectively calculate the position and posture information of the orthodontic teeth, obtain more accurate detection of the orthodontic effect, and can well consider the impact of the orthodontic on the expected effect, and adjust the orthodontic method in time. Strong industrial applicability.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Graphics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Dental Tools And Instruments Or Auxiliary Dental Instruments (AREA)

Abstract

本公开涉及一种牙齿矫正效果的检测方法、装置、设备和存储介质，牙齿矫正效果的检测方法包括：获取牙齿图像，牙齿图像中包括目标牙齿；基于预先训练的神经网络模型对牙齿图像中的目标牙齿进行识别，确定目标牙齿的特征信息；根据特征信息和目标牙齿对应的牙齿模型，确定目标牙齿的位姿信息；基于位姿信息将牙齿模型映射到牙齿图像中，并基于牙齿图像生成目标牙齿矫正效果的检测结果。本公开提供的方法，在牙齿诊疗过程中，不需要频繁复诊，通过采集牙齿图像就能够实时自动获取牙齿矫正效果的相关数据，对牙齿矫正效果进行检测，操作简便且便于实施，同时还便于后续相关医护人员了解牙齿矫正情况，及时调整矫正方案。

Description

牙齿矫正效果的检测方法、装置、设备和存储介质
本公开要求于2022年03月01日提交中国专利局、申请号为202210194371.2、发明名称为“牙齿矫正效果的检测方法、装置、设备和存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本公开中。
技术领域
本公开涉及信息处理技术领域，尤其涉及一种牙齿矫正效果的检测方法、装置、设备和存储介质。
背景技术
随着社会的发展，正畸治疗作为口腔美学治疗的一种方式越来越受到人们的重视。其中“三维数字化正畸”作为一种全新数字化矫正技术，被广泛应用。在口腔牙齿治疗过程中，一般是医生通过口内扫描仪的手柄采集用户的牙齿图像，然后查看显示器上显示的牙齿图像，进行诊断后判断出用户的牙齿缺失或缺损的情况。
传统的正畸治疗过程中，用户在每次复查时都需要通过口内扫描仪扫描用户的牙齿，来获取用户的当前牙齿模型并处理，然后与预设模型做比较，获取误差报告，获取治疗效果，但是，该过程比较耗时且复查成本高昂，会造成人力物力的浪费。
发明内容
(一)要解决的技术问题
本公开要解决的技术问题是解决现有的正畸治疗过程比较耗时且复查成本高昂的问题。
(二)技术方案
为了解决上述技术问题，本公开实施例提供了一种牙齿矫正效果的检测方法，包括：
获取牙齿图像，所述牙齿图像中包括目标牙齿；
基于预先训练的神经网络模型对所述牙齿图像中的目标牙齿进行识别，确定所述目标牙齿的特征信息；
根据所述特征信息和所述目标牙齿对应的牙齿模型，确定所述目标牙齿的位姿信息；
基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果。
第二方面，还提供一种牙齿矫正效果的检测装置，包括：
获取单元，用于获取牙齿图像，所述牙齿图像中包括目标牙齿；
识别单元，用于基于预先训练的神经网络模型对所述牙齿图像中的目标牙齿进行识别，确定所述目标牙齿的特征信息；
确定单元，用于根据所述特征信息和所述目标牙齿对应的牙齿模型，确定所述目标牙齿的位姿信息；
检测单元，用于基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果。
第三方面，还提供一种电子设备，包括：
存储器；
处理器；以及
计算机程序；
其中，所述计算机程序存储在所述存储器中，并被配置为由所述处理器执行以实现上述的牙齿矫正效果的检测方法。
第四方面，还提供一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述的牙齿矫正效果的检测方法的步骤。
(三)有益效果
本公开实施例提供的上述技术方案与现有技术相比具有如下优点：
本公开实施例提供的该牙齿矫正效果的检测方法，包括：获取牙齿图像；基于预先训练的神经网络模型对牙齿图像进行识别，确定牙齿图像中目标牙齿的特征信息；根据特征信息和目标牙齿对应的牙齿模型，确定目标牙齿的位姿信息；基于位姿信息将牙齿模型映射到牙齿图像中，并基于牙齿图像生成目标牙齿矫正效果的检测结果。本公开提供的方法，用户在牙齿诊疗过程中，只需要拍摄包括牙齿的二维图像即可，二维牙齿可以通过终端等具有拍摄功能的设备获取，不需要使用专业的口内扫描仪就可以让用户居家完成检测，所以不需要频繁复诊，可以只通过采集当前被矫正牙齿的图像就能够实时自动获取牙齿矫正效果的相关数据，对牙齿矫正效果进行检测，操作简便且便于实施，同时还便于后续相关医护人员了解矫正过程，及时调整矫正方案。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，并不能限制本公开。
附图说明
此处的附图被并入说明书中并构成本说明书的一部分，示出了符合本公开的实施例，并与说明书一起用于解释本公开的原理。
为了更清楚地说明本公开实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，对于本领域普通技术人员而言，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1为本公开实施例提供的一种牙齿矫正效果的检测方法的流程示意图；
图2为本公开实施例提供的一种应用场景的示意图；
图3为本公开实施例提供的一种牙齿图像的示意图；
图4为本公开实施例提供的一种牙齿模型的示意图；
图5为本公开实施例提供的一种牙齿矫正效果的检测方法的流程示意图；
图6为本公开实施例提供的一种牙齿矫正效果的检测装置的结构 示意图；
图7为本公开实施例提供的电子设备的结构示意图。
具体实施方式
为使本公开实施例的目的、技术方案和优点更加清楚，下面将对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本公开的一部分实施例，而不是全部的实施例。基于本公开中的实施例，本领域普通技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例，都属于本公开保护的范围。
随着社会的发展，正畸治疗作为口腔美学治疗的一种方式越来越受到人们的重视。正畸治疗主要通过对牙齿施加正畸力，引导牙周组织改建，从而改变牙齿在牙槽骨内的位置。正畸治疗可以改善由于牙列拥挤、牙齿异常排列等导致的咬合关系不良，从而达到牙周组织的长期稳定。“三维数字化正畸”作为一种全新数字化矫正技术，被广泛应用，对于错位牙、龅牙等矫正比传统正畸更为精准，其效果更自然、矫正时间更短、不反弹等优点而广受美齿者青睐。在口腔牙齿治疗过程中，一般是医生通过口内扫描仪的手柄采集用户的牙齿图像，然后查看显示器上显示的牙齿图像，进行诊断后判断出用户的牙齿缺失或缺损的情况。传统的正畸治疗过程中，用户需要多次来医院复诊，在每次复查时都需要通过口内扫描仪扫描用户的牙齿，来获取用户的当前牙齿模型并处理，然后与预设模型做比较，检查牙齿矫正情况是否与预先设计的方案匹配，以此来获取误差报告和治疗效果，但是，该过程耗时耗力，且复查成本高昂，用户配合度也不高，经常会出现延期复诊的现象，从而导致矫正方案失败率上升，影响最终治疗效果，对人力物力也是一种浪费。
针对上述技术问题，本公开实施例提供了一种牙齿矫正效果的检测方法，在牙齿诊疗过程中，通过实时拍摄牙齿图像，基于牙齿图像和预期模型对牙齿的当前矫正效果进行分析，操作简便的同时，还能为相关医护人员提供矫正数据，便于后续根据矫正数据进行调整，且 不需要频繁复查，一定程度上减少了人力物力的浪费。具体的，通过下述一个或多个实施例进行详细说明。
图1为本公开实施例提供的一种牙齿矫正效果的检测方法，用于对牙齿当前矫正效果进行检测，便于了解牙齿矫正情况，具体包括如图1所示的如下步骤S110至S140：
S110、获取牙齿图像，牙齿图像中包括目标牙齿。
具体的，牙齿矫正效果的检测方法可以由终端或服务器来执行。具体的，终端或服务器可以通过牙齿模型和牙齿图像对目标牙齿的矫正效果进行检测，目标牙齿是指单科牙齿，一个牙齿图像中可能包括多颗牙齿，对每颗牙齿进行检测时均可以称该颗牙齿为目标牙齿。例如，在一种应用场景中，如图2所示，服务器22包括牙齿模型，服务器22接收由终端21传输的牙齿图像，随后基于牙齿图像和牙齿模型对牙齿矫正效果进行检测。该牙齿图像可以是终端21拍摄获得的。或者，该牙齿图像是终端21从其他设备中获取的。再或者，该牙齿图像是终端21对预设图像进行图像处理后得到的图像，该预设图像可以是终端21拍摄获得的，或者该预设图像可以是终端21从其他设备中获取的。此处，并不对其他设备做具体限定。在另一种应用场景中，终端21包括牙齿模型，终端21获取牙齿图像后，根据牙齿图像和牙齿模型对牙齿矫正效果进行检测。
可以理解的是，本公开实施例提供的牙齿矫正效果的检测方法并不限于如上所述的几种可能场景。下面以终端21独自执行牙齿矫正效果的检测方法为例进行说明。
可理解的，终端通过采集装置获取包括目标牙齿的牙齿图像，牙齿图像中包括至少一颗比较完整的牙齿，若包括多颗牙齿则需要对多颗牙齿中的每颗牙齿的矫正情况进行分析，牙齿图像如图3所示，图3为本公开实施例提供的一种牙齿图像的示意图，图3中牙齿图像310包括两颗牙齿，两颗牙齿的部分轮廓清晰，两颗牙齿记为牙齿311和牙齿312；图3中牙齿图像320包括牙齿321。可理解的是，采集装置 可以内嵌到终端或者和终端组合使用的外置装置，终端可以是手机或平板电脑，内嵌到终端可以是终端本身带有的装置，例如终端配置的摄像头，摄像头拍摄牙齿的二维图像，或者拍摄一个牙齿的视频数据，该视频数据的每一帧就是一张二维的牙齿图像，外置装置可以是具有采集功能的装置，例如口腔拍摄仪器，口腔拍摄仪器也会拍摄口腔内牙齿的每一帧的图像，具体的外置装置不作限定。可理解的是，获取牙齿图像的方式不作限定，可根据用户需求自行选择。
可选的，终端在获取牙齿图像之前，需要获取牙齿模型，方法具体包括：获取牙齿扫描数据，并根据牙齿扫描数据构建三维牙齿模型，三维牙齿模型中包括至少一个牙齿对应的牙齿模型；牙齿扫描数据通过三维扫描仪获取。
可选的，将三维牙齿模型中每颗牙齿对应的牙齿模型进行编号。
可理解的，终端在拍摄牙齿图像之前，会通过三维扫描设备对牙齿进行扫描，得到牙齿扫描数据，例如某一用户的牙齿需要进行矫正，相关医护人员采用三维扫描仪对该用户的所有牙齿进行扫描，生成牙齿扫描数据，根据牙齿扫描数据构建用户牙齿矫正前的牙齿模型，所有单颗牙齿模型组成用户的三维牙齿模型，还可以对牙齿矫正前的每颗牙齿模型的位姿进行调整，得到矫正后的三维牙齿模型，矫正后的三维牙齿模型中所有牙齿模型组成预期模型，预期模型为矫正后最理想的效果模型，下述实施例以矫正后的牙齿模型为例进行说明。得到每颗牙齿矫正后的牙齿模型后，因为通过扫描数据得到的是整个三维牙齿模型，因此需要采用神经网络模型将该用户的三维牙齿模型(三维牙齿模型)划分为单颗牙齿模型，也就是说三维牙齿模型是由多个单颗牙齿模型组成的，并为每颗牙齿模型编号，编号便于对单颗牙齿模型进行定位，例如该用户包括28可牙齿，且每颗牙齿都需要进行矫正，对牙齿模型中28颗单颗牙齿模型进行编号，可以分别记为1至28。可理解的，还可以只对需要矫正的单颗牙齿模型进行编号，其余不需要矫正的牙齿则不进行编号。
示例性的，参见图4，图4为本公开实施例提供的一种牙齿模型的示意图，图4为三维牙齿模型，三维牙齿模型为三维模型，图4中牙齿模型411和牙齿模型412分别和图3的牙齿300中牙齿311和牙齿312对应，也就是说牙齿311对应的矫正后的牙齿模型是411，牙齿312对应的矫正后的牙齿模型为412，牙齿图像320中牙齿321对应的矫正后的牙齿模型为413。
S120、基于预先训练的神经网络模型对牙齿图像中的目标牙齿进行识别，确定目标牙齿的特征信息。
可选的，特征信息包括特征点、目标牙齿的坐标和目标牙齿的编号。
可理解的，在上述S110的基础上，终端获取牙齿图像后，基于预先训练的神经网络模型对牙齿图像进行识别，确定牙齿图像中目标牙齿的特征信息，预先训练的神经网络模型可以是具有特征提取和识别功能的模型，具体的模型不作限定，只要能够识别出牙齿得到特征信息即可；其中，特征信息包括牙齿图像的特征点、编号和坐标，若牙齿图像中包括多个牙齿，则识别出多个牙齿中每个牙齿的特征信息，也可以得到包括所有牙齿特征的一个特征信息，一个特征信息也可以理解为牙齿图像对应的特征信息，具体的，参见图3，图3中牙齿图像310包括牙齿311和牙齿312，牙齿图像320包括牙齿321，若针对单颗牙齿进行检测，牙齿图像310中牙齿311和牙齿312均可以记为目标牙齿，牙齿图像320中牙齿321记为目标牙齿，下述实施例以牙齿321为目标牙齿为例进行说明，牙齿321对应的特征信息包括牙齿321的特征点，特征点可以是牙齿图像上图像灰度值发生剧烈变化的点或者在牙齿图像边缘上曲率较大的点，牙齿321的坐标具体可以指牙齿321轮廓线(边缘线)包括的多个坐标点，牙齿321的编号用于在预先构建的三维牙齿模型中确定具体的单颗牙齿模型，也就是说牙齿321的编号和上述三维牙齿模型中某一单颗牙齿模型的编号相同，便于确定目标牙齿对应的牙齿模型，牙齿321对应的牙齿模型为图4中的412。
S130、根据特征信息和目标牙齿对应的牙齿模型，确定目标牙齿的位姿信息。
可理解的，在上述S120的基础上，根据目标牙齿对应的特征信息中的编号确定目标牙齿对应的牙齿模型，随后根据特征信息中的特征点和/或坐标以及牙齿模型的相关信息，确定目标牙齿的位姿信息，位姿信息包括朝向和空间位置信息，朝向是指牙齿的姿态，例如牙齿相对于三维牙齿模型来说是朝上或是朝下的姿态，空间位置信息也可以理解为三维坐标信息。
可选的，确定目标牙齿的位姿信息后，方法还包括：获取包括目标牙齿的其他牙齿图像，并根据其他牙齿图像对应的特征信息中的第一特征点和牙齿图像对应的特征信息中的第二特征点，更新位姿信息。
可理解的，若只获取到一张包括目标牙齿的图像，则直接根据特征信息中的坐标和目标牙齿对应的牙齿模型的相关信息计算目标牙齿的位姿信息；若获取到多张包括目标牙齿的图像，则根据多张牙齿图像中任一张牙齿图像对应的坐标和牙齿模型的相关信息计算得到位姿信息后，基于多张牙齿图像对应的特征点更新该位姿信息，也就是通过匹配不同牙齿图像中检测出的特征点对确定的位姿信息进行优化，提高准确度，多张牙齿图像对应的特征点中包括上述任一张牙齿图像包括的第一特征点，以及多张牙齿图像中除任一张牙齿图像的其余牙齿图像对应的第二特征点；或者，还可以根据多张牙齿图像计算的特征信息计算出目标牙齿的多个位姿信息，例如获取了包括牙齿321的10张牙齿图像，根据10张牙齿图像中每张牙齿图像对应的特征信息中的坐标分别计算出牙齿321对应的10个位姿信息，随后根据10张牙齿图像对应的特征信息中的特征点，联合优化上述确定的10个位姿信息，得到最优的位姿信息，以此来提高检测结果的准确率。
S140、基于位姿信息将牙齿模型映射到牙齿图像中，并基于牙齿图像生成目标牙齿矫正效果的检测结果。
可理解的，在上述S130的基础上，终端确定位姿信息后，根据位 姿信息将与牙齿图像中目标牙齿编号对应的牙齿模型投影到牙齿图像上，牙齿图像中只包括一个目标牙齿所得到的位姿信息也可以理解为牙齿图像对应的位姿信息，也就是建立牙齿图像中目标牙齿和牙齿模型之间的联系，进而建立牙齿图像的图像坐标系和牙齿模型的世界坐标系之间的联系；随后将牙齿模型投影到牙齿图像中，并基于映射后的牙齿图像生成目标牙齿矫正效果的检测结果，可理解的是，投影后的牙齿图像中同时包括二维的牙齿模型和目标牙齿，图像是二维图像，牙齿模型是三维牙齿模型，三维牙齿模型映射到二维牙齿图像后，牙齿图像中包括的就是二维牙齿模型，牙齿模型和目标牙齿存在大面积的重叠，但是在像素点上会存在差值，该差值可以反映矫正过程中牙齿的矫正情况，检测结果可以是矫正效果是否符合预期，例如参见图3，牙齿模型投影到牙齿图像320中得到映射图像330，映射图像330包括二维牙齿模型331和目标牙齿321，二维牙齿模型331将目标牙齿321的大部分覆盖。还可以计算牙齿图像和映射图像之间的像素差，以该牙齿图像中目标牙齿和映射图像中牙齿模型像素之间的差值，反映矫正过程中牙齿的矫正情况，检测结果可以是矫正效果是否符合预期。得到检测结果后可以将检测过程中涉及到的相关数据和检测结果发送至相关医护人员进行查验，由相关医护人员判断检测结果是否符合预期，便于后续快速了解矫正情况，及时调整矫正方式。
可选的，上述基于位姿信息将牙齿模型映射到牙齿图像中，并基于牙齿图像生成目标牙齿矫正效果的检测结果，具体包括：基于位姿信息将牙齿模型映射到牙齿图像中，得到映射图像；计算牙齿图像和映射图像之间的像素差；根据像素差生成目标牙齿矫正效果的检测结果。
可理解的，上述S140中生成检测结果具体包括如下实现流程：基于牙齿图像的位姿信息，将目标牙齿对应的牙齿模型映射到牙齿图像中，得到映射图像，参见图3，图3中牙齿图像320中包括牙齿321，得到牙齿321的位姿信息后，根据位姿信息将牙齿321对应的牙齿模 型413映射到牙齿图像320中，映射后的牙齿图像即映射图像记为330，映射图像330同时包括映射后的二维牙齿模型331以及被牙齿模型331遮盖大部分区域的目标牙齿321；得到映射图像330后，计算映射图像330中目标牙齿321和牙齿模型331之间的像素差，或者直接计算牙齿图像320和映射图像330之间的像素差，也就是计算目标牙齿321和牙齿模型331的拟合误差，得到像素差，最后对像素差进行分析，分析牙齿矫正是否符合预期，是否符合预期是指牙齿当前的矫正情况是否符合上述构建的矫正后的牙齿模型，即当前牙齿矫正的情况是否以牙齿模型为标准，还是出现了矫正误差，随后根据像素差和分析结果生成目标牙齿矫正效果的检测结果。
本公开实施例提供的一种牙齿矫正效果的检测方法，包括：获取牙齿图像，牙齿图像中包括目标牙齿；基于预先训练的神经网络模型对牙齿图像中目标牙齿进行识别，确定目标牙齿的特征信息；根据特征信息和目标牙齿对应的牙齿模型，确定目标牙齿的位姿信息；基于位姿信息将牙齿模型映射到牙齿图像中，并基于牙齿图像生成目标牙齿矫正效果的检测结果。本公开提供的方法，在牙齿诊疗过程中，用户可以根据移动终端自行采集牙齿图像，移动终端可以是手机，不需要使用专业的口内扫描仪就可以让用户居家完成检测，所以不需要用户频繁去复诊，就能实时便捷的获取牙齿矫正效果的相关数据，并通过实时自动的对比方法对牙齿矫正效果进行分析，生成检测结果，操作简便且便于实施，后续相关医护人员可以远程诊断，提高用户的医从性，也便于相关医护人员及时调整矫正方案，提升正畸矫正的成功率。
在上述实施例的基础上，可选的，根据特征信息和目标牙齿对应的牙齿模型，确定目标牙齿的位姿信息，具体包括如图5所示的如下步骤S510至S520：
S510、在三维牙齿模型中获取和目标牙齿编号相同的牙齿模型。
可理解的，终端根据牙齿扫描数据构建三维牙齿模型后，在三维 牙齿模型中选择和目标牙齿编号相同的牙齿模型，三维牙齿模型中包括用户待矫正牙齿的所有单颗牙齿模型，且单颗牙齿模型均存在编号，例如，图3中牙齿图像320中目标牙齿321的编号为16，在三维牙齿模型中获取编号同样为16的单颗牙齿模型，也就是上述图4中的牙齿模型413。
S520、根据特征信息和牙齿模型，确定目标牙齿的位姿信息。
可理解的，在上述S510的基础上，确定目标牙齿对应的牙齿模型后，根据特征信息中的坐标和牙齿模型的相关信息，确定目标牙齿的位姿信息。
可选的，根据特征信息和牙齿模型，确定目标牙齿的位姿信息，具体包括：确定牙齿模型的坐标系，并将目标牙齿的坐标转换到坐标系下；基于转换后的目标牙齿的坐标和牙齿模型的位姿信息，确定目标牙齿在坐标系下的位姿信息。
可理解的，根据特征信息和牙齿模型，确定目标牙齿的位姿信息，具体包括如下步骤：确定牙齿模型的坐标系，坐标系具体是世界坐标系，该坐标系是扫描牙齿的三维扫描仪确定的；确定坐标系后，将目标牙齿的坐标转换到该世界坐标系下，牙齿图像对应的坐标为图像坐标系，牙齿图像中目标牙齿的坐标也为图像坐标系，将目标牙齿的坐标转换到世界坐标系下也就是实现图像坐标系和世界坐标系之间的转换，也可以理解为将整个牙齿图像统一到世界坐标系，便于确定目标牙齿在世界坐标系下的位姿信息；随后基于转换后的目标牙齿的坐标和牙齿模型在世界坐标系下的位姿信息，确定目标牙齿在该世界坐标系下的位姿信息，目标牙齿的坐标具体可以指牙齿图像中组成目标牙齿轮廓线或者边缘线的多个坐标。
可选的，基于转换后的目标牙齿的坐标和牙齿模型的位姿信息，确定目标牙齿在坐标系下的位姿信息，具体包括：基于转换后的目标牙齿的坐标和牙齿模型的位置信息，将目标牙齿的轮廓线和牙齿模型的轮廓线进行拟合，确定目标牙齿的位置信息；根据牙齿模型的朝向， 确定目标牙齿的朝向；根据目标牙齿的朝向以及位置信息，得到目标牙齿在坐标系下的位姿信息。
可理解的，上述基于转换后的目标牙齿的坐标和牙齿模型的位姿信息，确定目标牙齿在坐标系下的位姿信息，具体包括如下步骤：位姿信息包括位置信息和朝向，位置信息为空间三维坐标，朝向用于确定牙齿在整个牙齿模型中的姿态；基于转换后的目标牙齿轮廓线的二维坐标和牙齿模型在世界坐标系下的位置信息，将目标牙齿的轮廓线和牙齿模型的轮廓线进行拟合，确定目标牙齿在该坐标系下的位置信息；随后根据牙齿模型的位姿信息中的朝向，确定目标牙齿在该坐标系下的朝向；最后根据目标牙齿的朝向和目标牙齿的位置信息，得到目标牙齿在坐标系下的位姿信息，确定目标牙齿的位姿信息后也就建立了牙齿图像和牙齿模型之间的转换关系，便于后续将牙齿模型映射到牙齿图像中，对矫正效果进行分析。
本公开实施例提供的一种牙齿矫正效果的检测方法，通过在三维牙齿模型中获取和目标牙齿编号相同的牙齿模型，也就是对牙齿模型进行定位确定目标牙齿对应的牙齿模型，随后根据特征信息中目标牙齿轮廓线的坐标和牙齿模型的位姿信息，确定目标牙齿在牙齿模型所在坐标系下的位姿信息，建立了牙齿模型和目标牙齿之间的联系，便于后续根据位姿信息将牙齿模型映射到牙齿图像中，对每颗牙齿的矫正效果进行分析。
图6为本公开实施例提供的牙齿矫正效果的检测装置的结构示意图。本公开实施例提供的牙齿矫正效果的检测装置可以执行上述牙齿矫正效果的检测方法实施例提供的处理流程，如图6所示，牙齿矫正效果的检测装置600包括：
获取单元610，用于获取牙齿图像，所述牙齿图像包括目标牙齿；
识别单元620，用于基于预先训练的神经网络模型对所述牙齿图像中目标牙齿进行识别，确定所述目标牙齿的特征信息；
确定单元630，用于根据所述特征信息和所述目标牙齿对应的牙齿 模型，确定所述目标牙齿的位姿信息；
检测单元640，用于基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果。
可选的，装置600还包括构建单元，构建单元用于所述获取牙齿图像之前，具体用于：
获取牙齿扫描数据，并根据所述牙齿扫描数据构建三维牙齿模型，所述三维牙齿模型中包括至少一个牙齿对应的牙齿模型；
所述牙齿扫描数据通过三维扫描仪获取。
可选的，装置600中所述特征信息包括所述目标牙齿的编号；所述三维牙齿模型中每颗牙齿对应的牙齿模型均存在编号。
可选的，确定单元630中所述根据所述特征信息和所述目标牙齿对应的牙齿模型，确定所述目标牙齿的位姿信息，具体用于：
在所述三维牙齿模型中获取和所述目标牙齿编号相同的牙齿模型；
根据所述特征信息和所述牙齿模型，确定所述目标牙齿的位姿信息。
可选的，装置600中所述特征信息还包括所述目标牙齿的坐标。
可选的，确定单元630中所述根据所述特征信息和所述牙齿模型，确定所述目标牙齿的位姿信息，具体用于：
确定所述牙齿模型的坐标系，并将所述目标牙齿的坐标转换到所述坐标系下；
基于转换后的所述目标牙齿的坐标和所述牙齿模型的位姿信息，确定所述目标牙齿在所述坐标系下的位姿信息。
可选的，装置600中所述位姿信息包括朝向和位置信息。
可选的，确定单元630中所述基于转换后的所述目标牙齿的坐标和所述牙齿模型的位姿信息，确定所述目标牙齿在所述坐标系下的位姿信息，具体用于：
基于转换后的所述目标牙齿的坐标和所述牙齿模型的位置信息，将所述目标牙齿的轮廓线和所述牙齿模型的轮廓线进行拟合，确定所述目标牙齿的位置信息；
根据所述牙齿模型的朝向，确定所述目标牙齿的朝向；
根据所述目标牙齿的朝向以及位置信息，得到所述目标牙齿在所述坐标系下的位姿信息。
可选的，装置600中所述特征信息还包括特征点，所述特征点用于标识所述目标牙齿的局部特征。
可选的，装置600中还包括更新单元，更新单元用于所述确定所述目标牙齿的位姿信息后，具体用于：
获取包括所述目标牙齿的其他牙齿图像，并根据所述其他牙齿图像对应的特征信息中的第一特征点和所述牙齿图像对应的特征信息中的第二特征点，更新所述位姿信息。
可选的，检测单元640中所述基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果，具体用于：
基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，得到映射图像；
计算牙齿图像和映射图像之间的像素差；
根据所述像素差生成所述目标牙齿矫正效果的检测结果。
图6所示实施例的牙齿矫正效果的检测装置可用于执行上述方法实施例的技术方案，其实现原理和技术效果类似，此处不再赘述。
图7为本公开实施例提供的一种电子设备的结构示意图。本公开实施例提供的电子设备可以执行上述实施例提供的牙齿矫正效果的检测方法处理流程，如图7所示，电子设备700包括：处理器710、通讯接口720和存储器730；其中，计算机程序存储在存储器730中，并被配置为由处理器710执行如上述的牙齿矫正效果的检测方法。
另外，本公开实施例还提供一种计算机可读存储介质，其上存储 有计算机程序，所述计算机程序被处理器执行以实现上述实施例所述的牙齿矫正效果的检测方法。
此外，本公开实施例还提供了一种计算机程序产品，该计算机程序产品包括计算机程序或指令，该计算机程序或指令被处理器执行时实现如上所述的牙齿矫正效果的检测方法。
此外，本公开实施例还提供了一种装置，该装置包括：
处理器；
用于存储处理器可执行指令的存储器；
其中，所述处理器被配置为：
获取牙齿图像，所述牙齿图像中包括目标牙齿；
基于预先训练的神经网络模型对所述牙齿图像中的目标牙齿进行识别，确定所述目标牙齿的特征信息；
根据所述特征信息和所述目标牙齿对应的牙齿模型，确定所述目标牙齿的位姿信息；
基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果。
需要说明的是，在本文中，诸如“第一”和“第二”等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
以上所述仅是本公开的具体实施方式，使本领域技术人员能够理解或实现本公开。对这些实施例的多种修改对本领域的技术人员来说将是显而易见的，本文中所定义的一般原理可以在不脱离本公开的精 神或范围的情况下，在其它实施例中实现。因此，本公开将不会被限制于本文所示的这些实施例，而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。
工业实用性
本公开提供的牙齿矫正效果的检测方法，可有效计算矫正牙齿的位姿信息，获得更准确的牙齿矫正效果检测，能够很好的考虑矫正牙齿对预期效果造成的影响，及时调整矫正方式，具有很强的工业实用性。

Claims (10)

  1. 一种牙齿矫正效果的检测方法，其特征在于，包括：
    获取牙齿图像，所述牙齿图像中包括目标牙齿；
    基于预先训练的神经网络模型对所述牙齿图像中的目标牙齿进行识别，确定所述目标牙齿的特征信息；
    根据所述特征信息和所述目标牙齿对应的牙齿模型，确定所述目标牙齿的位姿信息；
    基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果。
  2. 根据权利要求1所述的方法，其特征在于，所述获取牙齿图像之前，所述方法还包括：
    获取牙齿扫描数据，并根据所述牙齿扫描数据构建三维牙齿模型，所述三维牙齿模型中包括至少一个牙齿对应的牙齿模型；
    所述牙齿扫描数据通过三维扫描仪获取。
  3. 根据权利要求2所述的方法，其特征在于，所述特征信息包括所述目标牙齿的编号，所述三维牙齿模型中包括的每个牙齿模型均存在编号；所述根据所述特征信息和所述目标牙齿对应的牙齿模型，确定所述目标牙齿的位姿信息，包括：
    在所述三维牙齿模型中获取和所述目标牙齿编号相同的牙齿模型；
    根据所述特征信息和所述牙齿模型，确定所述目标牙齿的位姿信息。
  4. 根据权利要求3所述的方法，其特征在于，所述特征信息还包括所述目标牙齿的坐标；所述根据所述特征信息和所述牙齿模型，确定所述目标牙齿的位姿信息，包括：
    确定所述牙齿模型的坐标系，并将所述目标牙齿的坐标转换到所述坐标系下；
    基于转换后的所述目标牙齿的坐标和所述牙齿模型的位姿信息，确定所述目标牙齿在所述坐标系下的位姿信息。
  5. 根据权利要求4所述的方法，其特征在于，所述位姿信息包括朝向和位置信息；所述基于转换后的所述目标牙齿的坐标和所述牙齿模型的位姿信息，确定所述目标牙齿在所述坐标系下的位姿信息，包括：
    基于转换后的所述目标牙齿的坐标和所述牙齿模型的位置信息，将所述目标牙齿的轮廓线和所述牙齿模型的轮廓线进行拟合，确定所述目标牙齿的位置信息；
    根据所述牙齿模型的朝向，确定所述目标牙齿的朝向；
    根据所述目标牙齿的朝向以及位置信息，得到所述目标牙齿在所述坐标系下的位姿信息。
  6. 根据权利要求1所述的方法，其特征在于，所述特征信息还包括特征点，所述特征点用于标识所述目标牙齿的局部特征；
    所述确定所述目标牙齿的位姿信息后，所述方法还包括：
    获取包括所述目标牙齿的其他牙齿图像，并根据所述其他牙齿图像对应的特征信息中的第一特征点和所述牙齿图像对应的特征信息中的第二特征点，更新所述位姿信息。
  7. 根据权利要求1所述的方法，其特征在于，所述基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果，包括：
    基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，得到映射图像；
    计算所述牙齿图像和所述映射图像之间的像素差；
    根据所述像素差生成所述目标牙齿矫正效果的检测结果。
  8. 一种牙齿矫正效果的检测装置，其特征在于，包括：
    获取单元，用于获取牙齿图像，所述牙齿图像中包括目标牙齿；
    识别单元，用于基于预先训练的神经网络模型对所述牙齿图像中 的目标牙齿进行识别，确定所述目标牙齿的特征信息；
    确定单元，用于根据所述特征信息和所述目标牙齿对应的牙齿模型，确定所述目标牙齿的位姿信息；
    检测单元，用于基于所述位姿信息将所述牙齿模型映射到所述牙齿图像中，并基于所述牙齿图像生成所述目标牙齿矫正效果的检测结果。
  9. 一种电子设备，其特征在于，包括：
    存储器；
    处理器；以及
    计算机程序；
    其中，所述计算机程序存储在所述存储器中，并被配置为由所述处理器执行以实现如权利要求1至7中任一所述的牙齿矫正效果的检测方法。
  10. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至7中任一所述的牙齿矫正效果的检测方法的步骤。
PCT/CN2023/078959 2022-03-01 2023-03-01 牙齿矫正效果的检测方法、装置、设备和存储介质 WO2023165505A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210194371.2A CN114549509A (zh) 2022-03-01 2022-03-01 牙齿矫正效果的检测方法、装置、设备和存储介质
CN202210194371.2 2022-03-01

Publications (1)

Publication Number Publication Date
WO2023165505A1 true WO2023165505A1 (zh) 2023-09-07

Family

ID=81662379

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/078959 WO2023165505A1 (zh) 2022-03-01 2023-03-01 牙齿矫正效果的检测方法、装置、设备和存储介质

Country Status (2)

Country Link
CN (1) CN114549509A (zh)
WO (1) WO2023165505A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117857912A (zh) * 2024-01-15 2024-04-09 中山大学附属口腔医院 用于牙科的信息采集方法和系统

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114549509A (zh) * 2022-03-01 2022-05-27 先临三维科技股份有限公司 牙齿矫正效果的检测方法、装置、设备和存储介质
CN116593151B (zh) * 2023-07-17 2023-09-12 创新奇智(青岛)科技有限公司 牙套拉力器测试方法、装置、电子设备及可读存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108320325A (zh) * 2018-01-04 2018-07-24 华夏天宇（北京）科技发展有限公司 牙列模型的生成方法及装置
CN111784754A (zh) * 2020-07-06 2020-10-16 浙江得图网络有限公司 基于计算机视觉的牙齿正畸方法、装置、设备及存储介质
CN111931843A (zh) * 2020-08-10 2020-11-13 深圳爱舒笑科技有限公司 一种基于图像处理的监控牙齿位置的方法
CN112807108A (zh) * 2021-01-27 2021-05-18 清华大学 一种正畸矫治过程中的牙齿矫治状态的检测方法
CN114549509A (zh) * 2022-03-01 2022-05-27 先临三维科技股份有限公司 牙齿矫正效果的检测方法、装置、设备和存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108320325A (zh) * 2018-01-04 2018-07-24 华夏天宇（北京）科技发展有限公司 牙列模型的生成方法及装置
CN111784754A (zh) * 2020-07-06 2020-10-16 浙江得图网络有限公司 基于计算机视觉的牙齿正畸方法、装置、设备及存储介质
CN111931843A (zh) * 2020-08-10 2020-11-13 深圳爱舒笑科技有限公司 一种基于图像处理的监控牙齿位置的方法
CN112807108A (zh) * 2021-01-27 2021-05-18 清华大学 一种正畸矫治过程中的牙齿矫治状态的检测方法
CN114549509A (zh) * 2022-03-01 2022-05-27 先临三维科技股份有限公司 牙齿矫正效果的检测方法、装置、设备和存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117857912A (zh) * 2024-01-15 2024-04-09 中山大学附属口腔医院 用于牙科的信息采集方法和系统
CN117857912B (zh) * 2024-01-15 2024-06-07 中山大学附属口腔医院 用于牙科的信息采集方法和系统

Also Published As

Publication number Publication date
CN114549509A (zh) 2022-05-27

Similar Documents

Publication Publication Date Title
WO2023165505A1 (zh) 牙齿矫正效果的检测方法、装置、设备和存储介质
AU2021206867B2 (en) Automatic selection and locking of intraoral images
US11412922B2 (en) Control of light sources on an intraoral mirror with an integrated camera
US10893918B2 (en) Determining a dental treatment difficulty
Hutton et al. An evaluation of active shape models for the automatic identification of cephalometric landmarks
US12002271B2 (en) Identification device, scanner system, and identification method
JP2017529919A (ja) 口腔内デジタル三次元モデルを位置合わせする方法
JP7220569B2 (ja) 撮像支援装置、スキャナシステム、および撮像支援方法
CN110868913B (zh) 利用口内3d扫描图跟踪牙龈线并且显示牙周测量结果的工具
CN111931843A (zh) 一种基于图像处理的监控牙齿位置的方法
JP6547219B2 (ja) 歯科画像処理装置
US20230215027A1 (en) Oral image marker detection method, and oral image matching device and method using same
CN114830179A (zh) 确定上牙与面部骨骼之间的空间关系
Ahn et al. Marginal bone destructions in dental radiography using multi-template based on internet services

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23762904

Country of ref document: EP

Kind code of ref document: A1