CN115530762A - CBCT temporomandibular joint automatic positioning method and system - Google Patents
CBCT temporomandibular joint automatic positioning method and system Download PDFInfo
- Publication number
- CN115530762A CN115530762A CN202211241399.3A CN202211241399A CN115530762A CN 115530762 A CN115530762 A CN 115530762A CN 202211241399 A CN202211241399 A CN 202211241399A CN 115530762 A CN115530762 A CN 115530762A
- Authority
- CN
- China
- Prior art keywords
- cbct
- temporomandibular joint
- model
- training
- automatic positioning
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0033—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
- 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
- G06T2207/30204—Marker
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Radiology & Medical Imaging (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Data Mining & Analysis (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Computing Systems (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Quality & Reliability (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a CBCT temporomandibular joint automatic positioning method and system, belonging to temporomandibular joint positioning technical field, comprising the following steps: s1: acquiring a CBCT image; s2: img _ i slice i, img _ i is an image of a transverse, coronal, or sagittal plane; s3: and calculating the coordinates of the temporomandibular joint according to the index i and the frame. The S2 comprises the following steps: taking the coronal plane as an example, the starting index of i is k, and k is set according to the prior knowledge or k is set to 0; the positioning method and the positioning system realize real-time positioning through the target detection model, distinguish the left and right temporomandibular joints, realize the detection model on 2D data and have less computer resource requirement. The training samples are not markers for keypoints, and only a roughly rectangular box is required to contain the region of interest. The requirements on the experience of doctors are low, and the realization is easy.
Description
Technical Field
The invention relates to the technical field of temporomandibular joint positioning, in particular to a CBCT temporomandibular joint automatic positioning method and system.
Background
The temporomandibular joint (TMJ) is one of the most complex joints of the human body in terms of composition and function, and its structure and function are of great importance to the health of the oral-jaw system. Cone-beam CT (CBCT) has made significant progress in TMJ morphology studies compared to conventional X-ray films. Through CBCT, the doctor can not only obtain clear skeleton structure image fast, but also can directly carry out more accurate, accurate measurement to the position that needs the analysis.
In order to realize quantitative analysis of the temporomandibular joint bone structure, it is necessary to perform multi-planar visualization of the temporomandibular joint, such as transverse, coronal and sagittal planes. In the prior art, a doctor needs to manually select the position of a temporomandibular joint and is matched with a multi-plane reconstruction technology (MPR) to observe images at different angles, so that a focus is displayed better and more clearly. The manual selection of the temporomandibular joint position is inefficient and less intelligent, the positions selected by different doctors may be different, given quantitative analysis is more troublesome, and there is a lack of a competitive doctor who is prone to error.
Aiming at the problem that a manual positioning method is not used, a method and a device for positioning key points of a temporomandibular joint CBCT image are provided. According to the method, a large amount of computer resources are consumed when a 3D network is used, the coordinate points need to be marked manually by training samples, the requirement on marking precision is high, and the workload is large.
Disclosure of Invention
The invention is provided in view of the problems existing in the existing temporomandibular joint positioning process.
In order to solve the above technical problems, according to one aspect of the present invention, the present invention provides the following technical solutions: a CBCT temporomandibular joint automatic positioning method comprises the following steps:
s1: acquiring a CBCT image;
s2: img _ i slice i, img _ i is an image of a transverse, coronal, or sagittal plane;
s3: and calculating the coordinates of the temporomandibular joint according to the index i and the frame.
As a preferred embodiment of the CBCT temporomandibular joint automatic positioning method and system of the present invention, wherein: the S2 comprises the following steps:
taking the coronal plane as an example, the starting index of i is k, and k is set according to the prior knowledge or k is set to 0; the Img _ i is input into a pre-trained model to obtain a frame, a category and a confidence coefficient; when the confidence coefficient is greater than the set threshold th and is 0.9, the circulation is ended; otherwise i = i +1, the step is continued until all slices have been traversed.
As a preferred embodiment of the CBCT temporomandibular joint automatic positioning method and system of the present invention, wherein: the said S3, including,
the frame is represented by the coordinates of the lower left point (p 1_ x, p1_ z) and the upper right point (p 2_ x, p2_ z); point (x, y, z) = (0.5 × (p 1_ x + p2_ x), i,0.5 × (p 1_ z + p2_ z)); the categories distinguish the left and right temporomandibular joints.
A CBCT temporomandibular joint automatic positioning system comprises a setting network model, wherein the output end of the setting network model is connected with the input end of a training sample acquisition, and the output end of the training sample acquisition is connected with the input end of a pre-training model acquisition for model training.
As a preferable embodiment of the CBCT temporomandibular joint automatic positioning system of the present invention, wherein: the setting of the network model comprises model design and adopts a lightweight target detection model.
As a preferable aspect of the CBCT temporomandibular joint automatic positioning system of the present invention, wherein: the training sample acquisition comprises training sample preparation, CBCT image acquisition and frame marking.
As a preferable embodiment of the CBCT temporomandibular joint automatic positioning system of the present invention, wherein: the model training and pre-training model obtaining comprises the steps that training samples are input into a designed model, and training parameters are set to carry out network training by machine learning.
Compared with the prior art, the invention has the beneficial effects that:
the invention provides an artificial intelligence-based temporomandibular joint positioning method, which realizes real-time positioning and distinguishes left and right temporomandibular joints through a target detection model, and the detection model is realized on 2D data, so that the computer resource demand is low. The training sample is not a key point marker, and only a substantially rectangular box is required to contain the region of interest, as in the coronal plane marker example of FIG. 1, with other surface markers similar to the coronal plane. The method has the advantages of high efficiency, strong intellectualization and reliable repetition. The computer resource consumption is low, the CPU or the mobile terminal equipment can run fast and efficiently, and the doctor is helped to improve the diagnosis efficiency under the condition of not increasing extra cost. And the model of the scheme has low training cost and low requirements on the experience of a marking doctor, and is easy to popularize and realize.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail with reference to the accompanying drawings and detailed embodiments, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without inventive efforts. Wherein:
FIG. 1 is a first schematic representation of the present invention;
FIG. 2 is a second label schematic of the present invention;
FIG. 3 is a flow chart of the method of the present invention;
FIG. 4 is a block diagram of the system of the present invention;
FIG. 5 is a schematic diagram of the model predicted coordinates of the present invention.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways than those specifically described herein, and it will be apparent to those of ordinary skill in the art that the present invention may be practiced without departing from the spirit and scope of the present invention, and therefore the present invention is not limited to the specific embodiments disclosed below.
Examples
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
As shown in fig. 1-5, a CBCT temporomandibular joint automatic positioning method includes the following steps:
s1: the CBCT image is acquired.
S2: img _ i slice i, which may be a transverse, coronal, or sagittal image.
S3: calculating the coordinates of the temporomandibular joint according to the index i and the frame;
in this example, S2, includes:
taking the coronal plane as an example, the starting index of i is k, and k is set according to the prior knowledge (the position of the temporomandibular joint is basically in the eccentric area of the image volume data) or k is set to 0; the Img _ i is input into a pre-trained model to obtain a frame, a category and a confidence coefficient; when the confidence coefficient is larger than a set threshold th, such as 0.9, the loop is ended; otherwise i = i +1, the step is continued until all slices have been traversed.
In this example, S3, includes:
the frame is represented by the coordinates of the lower left point (p 1_ x, p1_ z) and the upper right point (p 2_ x, p2_ z); point (x, y, z) = (0.5 × (p 1_ x + p2_ x), i,0.5 × (p 1_ z + p2_ z)); the categories distinguish the left and right temporomandibular joints.
A CBCT temporomandibular joint automatic positioning system comprises a network model, wherein the output end of the network model is connected with the input end for obtaining training samples, and the output end for obtaining the training samples is connected with the input end for obtaining a pre-training model in model training.
In this example, setting up the network model includes model design, using a lightweight object detection model, such as yolov5, yolov7, or yolo-tiny. Taking yolov5s as an example, the model parameter is 7.5M, and the loss function includes 1) the classification loss cls _ loss: calculating whether the anchor frame and the corresponding calibration classification are correct, 2) positioning loss box _ loss: error between prediction and calibration boxes (GIoU), 3) confidence loss obj _ loss: a confidence level of the network is calculated.
In this example, acquiring the training sample includes training sample preparation, acquiring a CBCT image, and marking a border. And e.g. labeling the frame of the CBCT image by using tools such as LabelImg, ITK-SNAP and the like, wherein the frame is represented by coordinates of the lower left corner and the upper right corner. Frame marked on coronal plane as shown in FIG. 1
In this example, the model training to obtain the pre-training model includes inputting training samples into the designed model, and setting training parameters such as iteration times, learning rate, optimization method, and the like to perform network training. And adjusting model parameters through forward propagation and backward propagation of the network until the model converges, and acquiring a pre-trained model for deployment and reasoning.
As a result of the automatic positioning shown in FIG. 5, the method is efficient, highly intelligent, and reliable in repetition. The computer resource consumption is low, the CPU or the mobile terminal equipment can run fast and efficiently, and the doctor is helped to improve the diagnosis efficiency under the condition of not increasing extra cost. And the scheme model has low training cost, low requirements on the experience of a marking doctor and easy popularization and implementation.
While the invention has been described above with reference to an embodiment, various modifications may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In particular, the various features of the disclosed embodiments of the invention may be used in any combination, provided that no structural conflict exists, and the combinations are not exhaustively described in this specification merely for the sake of brevity and resource conservation. Therefore, it is intended that the invention not be limited to the particular embodiments disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.
Claims (7)
1. A CBCT temporomandibular joint automatic positioning method is characterized by comprising the following steps:
s1: acquiring a CBCT image;
s2: img _ i slice i, img _ i is an image of a transverse, coronal, or sagittal plane;
s3: and calculating the coordinates of the temporomandibular joint according to the index i and the frame.
2. The automatic CBCT temporomandibular joint positioning method of claim 1, wherein: the S2 comprises the following steps:
taking the coronal plane as an example, the starting index of i is k, and k is set according to the prior knowledge or k is set to 0; the Img _ i is input into a pre-trained model to obtain a frame, a category and a confidence coefficient; when the confidence coefficient is greater than the set threshold th and is 0.9, the circulation is ended; otherwise i = i +1, the step is continued until all slices have been traversed.
3. The automatic CBCT temporomandibular joint positioning method according to claim 1, wherein: the S3 comprises the following steps:
the frame is represented by the coordinates of the lower left point (p 1_ x, p1_ z) and the upper right point (p 2_ x, p2_ z); point (x, y, z) = (0.5 × (p 1_ x + p2_ x), i,0.5 × (p 1_ z + p2_ z)); the categories distinguish the left and right temporomandibular joints.
4. A CBCT temporomandibular joint automatic positioning system comprises a network model, and is characterized in that: the output end of the setting network model is connected with the input end of the training sample, and the output end of the training sample is connected with the input end of the pre-training model obtained by model training.
5. The CBCT temporomandibular joint automatic positioning system of claim 4, wherein: the setting of the network model comprises model design and adopts a lightweight target detection model.
6. The CBCT temporomandibular joint automatic positioning system of claim 4, wherein: the training sample acquisition comprises training sample preparation, CBCT image acquisition and frame marking.
7. The CBCT temporomandibular joint automatic positioning system of claim 4, wherein: the model training and pre-training model obtaining comprises the steps that training samples are input into a designed model, and training parameters are set to carry out network training by machine learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211241399.3A CN115530762A (en) | 2022-10-11 | 2022-10-11 | CBCT temporomandibular joint automatic positioning method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211241399.3A CN115530762A (en) | 2022-10-11 | 2022-10-11 | CBCT temporomandibular joint automatic positioning method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115530762A true CN115530762A (en) | 2022-12-30 |
Family
ID=84732662
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211241399.3A Pending CN115530762A (en) | 2022-10-11 | 2022-10-11 | CBCT temporomandibular joint automatic positioning method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115530762A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116152238A (en) * | 2023-04-18 | 2023-05-23 | 天津医科大学口腔医院 | Temporal-mandibular joint gap area automatic measurement method based on deep learning |
-
2022
- 2022-10-11 CN CN202211241399.3A patent/CN115530762A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116152238A (en) * | 2023-04-18 | 2023-05-23 | 天津医科大学口腔医院 | Temporal-mandibular joint gap area automatic measurement method based on deep learning |
| CN116152238B (en) * | 2023-04-18 | 2023-07-18 | 天津医科大学口腔医院 | Temporal-mandibular joint gap area automatic measurement method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107767376B (en) | X-ray bone age prediction method and system based on deep learning | |
| CN102422307B (en) | For method, system, device and computer program that interactive liver vessel and biliary system are assessed | |
| Shi et al. | An automatic method of fish length estimation using underwater stereo system based on LabVIEW | |
| CN107358608B (en) | Automatic bone tissue geometric parameter measuring device and method based on image processing technology | |
| CN109141248A (en) | Pig weight measuring method and system based on image | |
| CN108961330B (en) | Pig body length measuring and calculating method and system based on image | |
| CN101542532B (en) | A method, an apparatus and a computer program for data processing | |
| CN109544536A (en) | The fast automatic analysis method of hip joint x-ray image | |
| CN110530477B (en) | Replacement gilt weight estimation method | |
| CN110782467B (en) | Horse body ruler measuring method based on deep learning and image processing | |
| CN110458817A (en) | Qualitative forecasting method, device, equipment and the storage medium of medical image | |
| CN104143193B (en) | Target object dividing method, apparatus and system in a kind of ultrasonoscopy | |
| Zhang et al. | Development and validation of a visual image analysis for monitoring the body size of sheep | |
| CN107945156A (en) | A kind of method of automatic Evaluation numeral pathology scan image image quality | |
| CN110766694B (en) | Interactive segmentation method of three-dimensional medical image | |
| CN1864074A (en) | Determining patient-related information on the position and orientation of mr images by the individualisation of a body model | |
| CN115530762A (en) | CBCT temporomandibular joint automatic positioning method and system | |
| CN114066872A (en) | Self-iteration wound evaluation system based on deep camera shooting and deep learning | |
| US20080285831A1 (en) | Automatically updating a geometric model | |
| Daemen et al. | Photographic documentation and severity quantification of pectus excavatum through three-dimensional optical surface imaging | |
| CN113706512A (en) | Live pig weight measurement method based on deep learning and depth camera | |
| CN111986216A (en) | RSG liver CT image interactive segmentation algorithm based on neural network improvement | |
| CN113643401B (en) | Live pig carcass segmentation method and system based on machine learning | |
| CN111144506A (en) | Ultrasound image-based hydatid identification method, storage medium and ultrasound equipment | |
| Lynn et al. | Automatic assessing body condition score from digital images by active shape model and multiple regression technique |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |