WO2023008849A1 - 3차원 치아 이미지 표시 장치 및 방법, 그리고 3차원 치아 이미지 표시 장치를 학습시키는 방법 - Google Patents
3차원 치아 이미지 표시 장치 및 방법, 그리고 3차원 치아 이미지 표시 장치를 학습시키는 방법 Download PDFInfo
- Publication number
- WO2023008849A1 WO2023008849A1 PCT/KR2022/010849 KR2022010849W WO2023008849A1 WO 2023008849 A1 WO2023008849 A1 WO 2023008849A1 KR 2022010849 W KR2022010849 W KR 2022010849W WO 2023008849 A1 WO2023008849 A1 WO 2023008849A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- tooth
- image
- dental
- dimensional
- learning
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000001514 detection method Methods 0.000 claims abstract description 44
- 238000013528 artificial neural network Methods 0.000 claims description 65
- 230000011218 segmentation Effects 0.000 claims description 39
- 238000007408 cone-beam computed tomography Methods 0.000 claims description 15
- 239000003086 colorant Substances 0.000 claims description 7
- 230000000873 masking effect Effects 0.000 claims description 4
- 239000005557 antagonist Substances 0.000 claims 2
- 238000003709 image segmentation Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 5
- 230000003042 antagnostic effect Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 230000014509 gene expression Effects 0.000 description 2
- 238000002513 implantation Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 239000004053 dental implant Substances 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000007943 implant Substances 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000000214 mouth Anatomy 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61C—DENTISTRY; APPARATUS OR METHODS FOR ORAL OR DENTAL HYGIENE
- A61C9/00—Impression cups, i.e. impression trays; Impression methods
- A61C9/004—Means or methods for taking digitized impressions
- A61C9/0046—Data acquisition means or methods
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61C—DENTISTRY; APPARATUS OR METHODS FOR ORAL OR DENTAL HYGIENE
- A61C13/00—Dental prostheses; Making same
- A61C13/0003—Making bridge-work, inlays, implants or the like
- A61C13/0004—Computer-assisted sizing or machining of dental prostheses
-
- 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
-
- 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
- G06T1/00—General purpose image data processing
- G06T1/0007—Image acquisition
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- 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
Definitions
- the present invention relates to a 3D tooth image display device and method, and a method for learning a 3D tooth image display device, and displays a 3D tooth image from a dental cross-sectional image in which an oral region is captured using an artificial neural network and displays a tooth number. It relates to a 3D tooth image display device and method for detecting, and a method for learning a 3D tooth image display device.
- a dental implant number which is a tooth placement number of a subject for implantation, is required.
- a medical practitioner manually checks a dental formula number using a two-dimensional, low-dose conebeam computed tomography (CBCT) image.
- CBCT conebeam computed tomography
- An object of the present invention for solving the above problems is a 3D tooth image display device, display method, and method for displaying a 3D tooth image accurately classified by tooth formula number while reducing data throughput and loading time, and displaying the same efficiently and quickly. It is about providing a way to learn.
- a 3D tooth image display device for achieving the above object is to divide an input 3D dental image in an axial direction to determine the boundary of each object in a dental cross-sectional image generated.
- a tooth image acquisition unit generating a 3D image for each tooth based on the boundary of each object in the image and the tooth number of each tooth, and the 3D image for each tooth may be displayed in different ways according to the tooth number.
- the tooth segmentation unit includes a tooth segmentation neural network, and the tooth segmentation neural network uses a plurality of dental cross-sectional images for learning as input data for learning, and the masking image in which the boundary area of each object in the learning dental cross-sectional image is masked is used as label data. can be learned using
- the tooth detection unit includes a tooth detection neural network, and the tooth detection neural network uses a plurality of dental cross-sectional images for learning as input data for learning, a tooth region in which a tooth exists in the dental cross-sectional image for learning, and each tooth in the tooth region. It can be learned using the tooth number for the as label data.
- the tooth image acquisition unit includes a tooth image acquisition neural network, and the tooth image acquisition neural network includes a dental cross-sectional image for learning in which a three-dimensional image for learning is divided along one axis, boundary information of each tooth in each dental cross-sectional image for learning, and a tooth for each tooth. Based on the number, it can be learned to create each 3D image for each tooth having the same tooth number in the 3D image.
- the tooth image acquisition neural network performs segmented images for each tooth having the same tooth number for the segmented images for each tooth extracted according to the boundary information within each segmented learned dental cross-sectional image and the extracted segmented images for each tooth. It may be learned to generate the respective 3D images by stacking them based on the uniaxial direction.
- the display method may include displaying at least some of the teeth in different colors according to the tooth number for each tooth.
- the display method includes displaying in the same color when the tooth number relationship for each tooth is the same name or antagonistic relationship, and displaying in different colors when the tooth number relationship for each tooth is not the same name or antagonistic relationship. can do.
- the 3D dental image may include a CBCT image, and the segmented dental cross-sectional image may be an image perpendicular to the axial direction.
- a method for displaying a 3D tooth image according to another embodiment of the present invention for achieving the above object is to determine the boundary of each object in a dental cross-sectional image generated by dividing an input 3D dental image in an axial direction.
- Based on, generating a 3D image for each tooth, and the 3D image for each tooth may have different display methods according to tooth numbers.
- a method for learning a 3D tooth image segmentation device divides an input 3D dental image in an axial direction and determines the boundary of each object in a dental cross-sectional image generated.
- Learning a tooth segmentation neural network to segment learning a tooth detection neural network to recognize a tooth region among objects in the generated dental cross-sectional image and detecting a tooth number of each tooth belonging to the tooth region, and the tooth Learning a tooth acquisition neural network to generate a three-dimensional image for each tooth based on the tooth number of each tooth and the boundary of each object in each dental cross-sectional image acquired through the segmentation neural network and the tooth detection neural network.
- the dental cross-sectional image for learning is a three-dimensional learning image divided along one axis, and the step of learning the tooth image acquisition neural network is based on the boundary information of each tooth in each dental cross-sectional image for learning and the tooth number for each tooth. , learning to generate each 3D image for each tooth having the same tooth number in the 3D image.
- a 3D tooth image display device and method according to an embodiment of the present invention, and a method for learning a 3D tooth image display device learn an artificial neural network model to recognize at least one tooth number from a dental cross-sectional image in an axial direction. do.
- Apparatus and method according to an embodiment of the present invention provides a three-dimensional image divided by tooth, compared to a conventional tooth detection method learned to detect and divide teeth in the form of a plurality of three-dimensional cells using a three-dimensional CBCT image Data learning and data overhead problems that occur when recognizing and displaying teeth and tooth formula numbers in an arbitrary patient's tooth image are prevented.
- a 3D tooth image display device display method, and method for displaying a 3D tooth image accurately classified by tooth formula number and learning them efficiently and quickly It can be provided.
- FIG. 1 is a block diagram for explaining a three-dimensional tooth image display device according to an embodiment of the present invention.
- Figure 2 is an output result image of the tooth segmentation unit of the three-dimensional tooth image display device according to an embodiment of the present invention.
- Figure 3 is an output result image of the tooth detection unit of the three-dimensional tooth image display device according to an embodiment of the present invention.
- Figure 4 is an output result image of the image segmentation unit of the three-dimensional tooth image display device according to an embodiment of the present invention.
- 5 is a final output result image of the three-dimensional tooth image display device according to an embodiment of the present invention.
- Figure 6 is a block diagram for explaining the three-dimensional tooth image display device according to an embodiment of the present invention in terms of hardware.
- FIG. 7 is a flowchart of a 3D tooth image learning method according to another embodiment of the present invention.
- first element may be termed a second element, and similarly, a second element may be termed a first element, without departing from the scope of the present invention.
- the term “and/or” includes any combination of a plurality of related listed items or any of a plurality of related listed items.
- FIG. 1 is a block configuration diagram for explaining a three-dimensional tooth image segmentation display device according to an embodiment of the present invention.
- the 3D tooth image display device 1000 includes a data receiving unit 1100, an image segmentation unit 1200, a tooth segmentation unit 1300, a tooth detection unit 1400, a tooth image acquisition unit 1500, and A display unit 1600 may be included.
- the 3D tooth image display apparatus 1000 may detect a boundary of an object included in each of a plurality of dental cross-sectional images using a pretrained tooth segmentation neural network model.
- the 3D tooth image display device 1000 may detect the width and number of at least one tooth by using a tooth detection neural network model in a dental cross-sectional image.
- the 3D tooth image display device 1000 may output the detected at least one tooth as a 3D image. Therefore, the 3D tooth image display device 1000 can provide tooth placement information necessary for implantation.
- the data receiving unit 1100 may receive at least one 3D dental image in which at least a part of the oral cavity region is photographed from the outside.
- the data receiving unit 1100 may receive at least one 3D dental image from an external photographing device.
- a 3D dental image input from an external imaging device may include a 3D CBCT dental image.
- the image segmentation unit 1200 may generate a plurality of dental cross-sectional images by dividing the three-dimensional dental image received from the data receiver 1100 into cross-sectional images perpendicular to an axial direction.
- the image segmentation unit 1200 may generate a plurality of dental cross-sectional images by slicing the 3D CBCT dental image vertically in an axial direction.
- the dental cross-sectional image may be an axial cross-sectional image perpendicular to the axial direction in a CBCT image including gray scale information having a range of 0 or more to 255 or less.
- the tooth segmentation unit 1300 may include a tooth segmentation neural network 1350 .
- the tooth segmentation unit 1300 receives at least one dental cross-sectional image from the image segmentation unit 1200, and outputs a first image in which at least one object included in the received dental cross-sectional image is masked. there is.
- the tooth segmentation unit 1300 may input at least one dental cross-section image received from the image segmentation unit 1200 to the tooth segmentation neural network 1350 . Also, the tooth segmentation unit 1300 may detect a boundary of at least one object included in the dental cross-sectional image by using the tooth segmentation neural network 1350 .
- the tooth dividing unit 1300 may generate a first image in which the boundary of at least one object is marked and individually masked.
- an object for which the tooth segmentation unit 1300 detects a boundary using the tooth segmentation neural network 1350 may include at least one tooth.
- the tooth segmentation neural network 1350 may be a machine-learned model using at least one dental cross-sectional image in which at least one tooth exists and a masking image mapped with CT coordinates of teeth in the dental cross-sectional image for learning as label data.
- the tooth segmentation neural network 1350 may include a convolutional neural network.
- the tooth segmentation unit 1300 may use the tooth segmentation neural network 1350 for any received dental cross-sectional image to identify tooth boundaries included in the dental cross-sectional image, and obtain a first image in which the tooth boundaries are marked. .
- Figure 2 is a resultant image of the tooth division of the three-dimensional tooth image display device according to an embodiment of the present invention.
- the tooth segmentation unit 1300 may receive at least one dental cross-section image 20 from the image segmentation unit 1200 . Subsequently, the tooth segmentation unit 1300 may obtain a masked tooth image 22 by dividing a boundary of at least one tooth in the received dental cross-sectional image 20 using the tooth segmentation neural network 1350.
- the tooth detection unit 1400 may include a tooth detection neural network 1450 .
- the tooth detecting unit 1400 may generate attribute information for each tooth included in the dental cross-sectional image from at least one dental cross-sectional image received from the image dividing unit 1200 .
- the tooth detection unit 1400 may input at least one dental cross-section image received from the image segmentation unit 1200 to the tooth detection neural network 1450 .
- the tooth detector 1400 represents attribute information for each tooth including at least one of a tooth width and a tooth number from at least one tooth included in the dental cross-sectional image by using the tooth detection neural network 1450.
- a second image may be obtained.
- the tooth-specific attribute information for tooth number 22 may include the width of tooth number 22, the tooth number of tooth number 22, or both.
- the tooth detection neural network 1450 may be trained using a convolutional neural network.
- the tooth detection neural network 1450 may be machine-learned using at least one dental cross-sectional image in which at least one tooth exists and a tooth number and location mapped to CT coordinates of teeth in the dental cross-sectional image as label data. .
- the tooth detecting unit 1400 may input the dental cross-section image into the tooth detecting neural network 1450 to obtain a second image marked with attribute information for each tooth including a tooth number.
- Figure 3 is an output result image of the tooth detection unit of the three-dimensional tooth image display device according to an embodiment of the present invention.
- the tooth detection unit 1400 detects the detection information included in the second image on the dental cross-section image 20 received from the data reception unit 1100 in a predetermined form for each tooth. can be obtained by For example, the tooth detection unit 1400 may detect detection information in the form of a box 32 .
- the tooth detector 1400 may obtain at least one attribute information for each tooth from coordinate data acquired using the tooth detection neural network 1450 .
- the attribute information may include at least one of position and width information for each tooth.
- the tooth image acquisition unit 1500 divides at least one first image and a second image obtained from the tooth division unit 1300 and the tooth detection unit 1400 for each tooth, respectively.
- a plurality of 2D segmented images may be acquired.
- the tooth image acquisition unit 1500 may classify the plurality of 2D divided images for each tooth based on at least one piece of information acquired from the tooth division unit 1300 and the tooth detection unit 1400 .
- the at least one piece of information may be at least one of a tooth number, tooth coordinate data, and boundary information between teeth.
- the tooth image acquisition unit 1500 may include a tooth image acquisition neural network 1550 . And, the tooth image acquiring unit 1500 may obtain a third image by reconstructing at least one 2D split image of at least one first image and at least one second image classified as the same tooth in 3D.
- the tooth image obtaining unit 1500 may obtain a third image, which is an individual tooth image reconstructed in 3D, by stacking 2D segmented images classified as the same tooth in an axial direction.
- the tooth image acquisition neural network 1550 is based on the dental cross-sectional image for learning in which the 3D image is divided along one axis, the boundary information of each tooth in each dental cross-sectional image for learning, and the tooth number for each tooth in advance. , it can be learned to generate each 3D image for each tooth having the same tooth number in the 3D image.
- the tooth image acquisition neural network 1550 converts the divided images for each tooth having the same tooth number for the divided images for each tooth extracted according to the boundary information within each divided learning dental cross-sectional image and the extracted divided images for each tooth in the uniaxial direction. It is possible to create a three-dimensional image for each tooth by stacking based on.
- the tooth image acquisition neural network 1550 may be trained to generate each 3D image for each tooth having the same tooth number in the 3D image using the stacked images.
- the tooth image obtaining unit 1500 may mask at least some of the teeth with different colors according to the number or type of teeth in consideration of the number or type of teeth present in the 3D dental image.
- the tooth image obtaining unit 1500 may mask with the same color when the tooth number relationship for each tooth is the same or antagonistic.
- the tooth image obtaining unit 1500 may mask each other with different colors when the tooth number relationship for each tooth is not the same or antagonistic.
- the 3D tooth image display device 1000 acquires a 3D image for each tooth based on the first image and the second image, which are 2D images, thereby reducing the amount of data processing and loading time. It is possible to provide a 3D dental image that is accurately classified for each part.
- Figure 4 is an output result image of the tooth image acquisition unit of the three-dimensional tooth image display device according to an embodiment of the present invention.
- the tooth image acquisition unit 1500 receives the first image and the second image from the tooth dividing unit 1300 and the tooth detection unit 1400, respectively, and creates a three-dimensional image 42 for each tooth. can create
- the tooth image acquisition unit 1500 may generate a 3D dental image masked at coordinates to which the 3D image for each tooth is mapped using the tooth image acquisition neural network 1550 .
- the tooth image obtaining unit 1500 may generate a 3D dental image in which the 3D image 42 for each tooth is set so that the display method is different according to the tooth number. Also, the display unit 1600 may output the 3D image 42 for each tooth in different colors according to the tooth number for each tooth.
- the tooth image obtaining unit 1500 may set the 3D image 42 for each tooth to be displayed in red for teeth 11, 21, 31, and 41 of the same type.
- the display unit 1600 may display the 3D image 42 for each tooth displayed in red.
- 5 is a final output result image of the three-dimensional tooth image display device according to an embodiment of the present invention.
- the 3D tooth image display apparatus 1000 may output a 3D tooth region in which individual teeth are separated and displayed as a final output result image.
- the 3D tooth image display apparatus 1000 may display a region where teeth are displayed separately from other regions.
- FIG. 6 the hardware configuration of the three-dimensional tooth image display device according to an embodiment of the present invention will be described in more detail.
- Figure 6 is a block diagram for explaining the hardware configuration of the three-dimensional tooth image display device according to an embodiment of the present invention.
- the three-dimensional tooth image display device 1000 includes a storage device 1210 for storing at least one command and a processor 1220 for executing at least one command of the memory, transmitting and receiving. device 1230 , an input interface device 1240 and an output interface device 1250 .
- Each of the components 1210, 1220, 1230, 1240, 1250 included in the 3D tooth image display device 1000 are connected by a data bus 1260 to communicate with each other.
- the storage device 1210 may include at least one of a memory or a volatile storage medium and a non-volatile storage medium.
- the storage device 1210 may include at least one of a read only memory (ROM) and a random access memory (RAM).
- the storage device 1210 may further include at least one command to be executed by the processor 1220 to be described later.
- At least one command is a first command for learning the tooth segmentation neural network 1350 to divide the boundary of each object in the dental cross-section image segmented in the axial direction, the dental cross-section for learning and a second command to train the tooth detection neural network 1450 to detect the tooth region and tooth number of each tooth in the image.
- a third command for training the tooth image acquisition neural network 1550 to output 3D segmented images for each of a plurality of teeth reconstructed into a 3D image including the tooth number and the tooth position may be included.
- the processor 1220 performs image segmentation unit 1200, tooth segmentation unit 1300, and tooth detection unit 1400 by at least one program command stored in the storage device 1210. And it can perform the functions of the tooth image acquisition unit 1500, each of which can be stored in a memory in the form of at least one module and executed by a processor.
- the input interface device 1240 may receive at least one control signal from a user, and the output interface device 1250 may output at least one piece of information related to the 3D tooth image by the operation of the processor 1220 to visualize it.
- the transmitting and receiving device 1230 in the 3D tooth image display device 1000 may receive at least one 3D dental image in which teeth are photographed from the outside (S1000). ).
- the processor 1220 may obtain a dental cross-sectional image by dividing the 3D dental image in an axial direction (S2000).
- the 3D dental image may be a 3D CBCT image.
- the dental cross-sectional image may be a cross-sectional image obtained by slicing a CBCT image in an axial direction.
- the processor 1220 may detect at least one tooth boundary from at least one dental cross-sectional image by using the pretrained tooth segmentation neural network 1350 .
- the processor 1220 may obtain a first image obtained by dividing a boundary for each tooth from at least one dental cross-sectional image using the tooth segmentation neural network 1350 (S3000).
- the tooth segmentation neural network 1350 uses a plurality of dental cross-sectional images for learning as input data for learning, and uses a masking answer image in which the correct answer of each object in the dental cross-sectional image for learning is masked as label data. Can be learned there is.
- the tooth segmentation neural network 1350 may be trained using a convolutional neural network.
- the tooth detection neural network 1450 uses a plurality of dental cross-sectional images for learning as input data for learning, and uses a tooth region where a tooth exists in the dental cross-sectional image for learning and a tooth number for each tooth in the tooth region as label data. can be learned using Also, the tooth detection neural network 1450 may be trained using a convolutional neural network.
- the processor 1220 may output coordinate data within a plurality of dental cross-sectional images, and may obtain a second image including attribute information for each tooth in the dental cross-sectional image using the tooth detection neural network 1450.
- the attribute information may include at least one of a width for each tooth and the number of teeth.
- the processor 1220 may obtain a 3D third image including the region of interest from at least one of the first image and the second image.
- the processor 1220 divides the first image and the second image by tooth, classifies the divided images by the same tooth using the pretrained tooth image acquisition neural network 1550, and stacks them along the axial direction.
- the processor 1220 may obtain a 3D tooth image reconstructed in 3D (S5000).
- the processor 1220 may output a 3D image of teeth through the output interface device 1250 (S5000).
- the 3D tooth image learning method can shorten the implant procedure time, remove noise generated by crown deformation of the tooth, and convert the 3D section image of the center of the tooth to may be usable.
- Combinations of each block of the block diagram and each step of the flowchart attached to the present invention may be implemented by a computer program or instruction codes.
- a computer program or code may create means for performing the functions described in each block or each step of the flowchart. Since these computer programs or codes may be stored in any kind of computer usable or computer readable recording medium, the programs or codes stored in the recording medium, etc. may produce a block or manufactured item that performs the function of each step of the flowchart. . In addition, these programs or codes may be loaded on a computer or other programmable data processing equipment, so that the computer or other programmable data processing equipment may perform a series of operation steps.
- Each block or step may represent a module or portion of code comprising one or more executable programs or codes, and it is possible for blocks or steps in a flowchart to be performed out of sequence. For example, two blocks or steps shown in succession may be performed concurrently, or the blocks or steps may be performed in reverse order depending on their function.
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Public Health (AREA)
- Epidemiology (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Primary Health Care (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Dentistry (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Radiology & Medical Imaging (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Processing (AREA)
- Dental Tools And Instruments Or Auxiliary Dental Instruments (AREA)
- Image Analysis (AREA)
Abstract
Description
Claims (12)
- 입력된 3차원 치과 이미지에 대해 액시얼(Axial) 방향으로 분할하여 생성된 치과 단면 이미지 내에서 각 객체의 경계를 분할하는 치아 분할부;상기 치과 단면 이미지 내의 객체중에서 치아 영역을 인식하고, 상기 치아 영역에 속하는 치아들 각각의 치아 번호를 검출하는 치아 검출부; 및상기 치과 단면 이미지 내의 각 객체의 경계 및 상기 치아들 각각의 치아 번호를 기초로, 치아별 3차원 이미지를 생성하는 치아 이미지 획득부를 포함하고,상기 치아별 3차원 이미지는 치아 번호에 따라 표시 방법이 상이한 3차원 치아 이미지 표시 장치.
- 제1 항에 있어서,상기 치아 분할부는 치아 분할 신경망을 포함하고,상기 치아 분할 신경망은,복수의 학습용 치과 단면 이미지를 학습용 입력 데이터로서 이용하고, 상기 학습용 치과 단면 이미지 내의 각 객체의 경계 영역이 마스킹된 마스킹 이미지를 레이블 데이터로서 이용하여 학습되는 3차원 치아 이미지 표시 장치.
- 제1 항에 있어서,상기 치아 검출부는 치아 검출 신경망을 포함하고,상기 치아 검출 신경망은,복수의 학습용 치과 단면 이미지를 학습용 입력 데이터로서 이용하고, 상기 학습용 치과 단면 이미지 내의 치아가 존재하는 치아 영역 및 상기 치아 영역 내의 각 치아에 대한 치아번호를 레이블 데이터로서 이용하여 학습되는 3차원 치아 이미지 표시 장치.
- 제1 항에 있어서,상기 치아 이미지 획득부는, 치아 이미지 획득 신경망을 포함하고,상기 치아 이미지 획득 신경망은,학습용 3차원 이미지가 일축을 따라 분할된 학습용 치과 단면 이미지, 각 학습용 치과 단면 이미지 내의 각 치아의 경계 정보 및 각 치아별 치아번호를 기초로, 상기 3차원 이미지 내의 동일한 치아번호를 갖는 각 치아별로 각각의 3차원 이미지를 생성하도록 학습되는 3차원 치아 이미지 표시 장치.
- 제4 항에 있어서,상기 치아 이미지 획득 신경망은,상기 각 분할된 학습된 치과 단면 이미지 내에서 상기 경계 정보에 따라 추출된 치아별 분할 이미지들 및 추출된 상기 치아별 분할 이미지들에 대해 동일한 치아번호를 갖는 치아별 분할 이미지들을 상기 일축 방향을 기준으로 적층하여 상기 각각의 3차원 이미지를 생성하도록 학습되는 3차원 치아 이미지 표시 장치.
- 제1 항에 있어서,상기 표시 방법은 각 치아별 치아번호에 따라 적어도 일부는 서로 다른 색상으로 디스플레이되는 것을 포함하는 3차원 치아 이미지 표시 장치.
- 제6 항에 있어서,상기 표시 방법은 각 치아별 치아번호 관계가 동명치, 대합치 관계인 경우, 같은 색상으로 디스플레이되고, 각 치아별 치아번호 관계가 동명치, 대합치 관계가 아닌 경우, 서로 다른 색상으로 디스플레이되는 것을 포함하는 3차원 치아 이미지 표시 장치.
- 제1 항에 있어서,상기 3차원 치과 이미지는 CBCT 이미지를 포함하고, 상기 분할된 치과 단면 이미지는 상기 액시얼 방향에 수직한 이미지인, 3차원 치아 이미지 표시 장치.
- 입력된 3차원 치과 이미지에 대해 액시얼(Axial) 방향으로 분할하여 생성된 치과 단면 이미지 내에서 각 객체의 경계를 분할하는 단계와,상기 치과 단면 이미지 내의 객체중에서 치아 영역을 인식하고, 상기 치아 영역에 속하는 각 치아의 치아 번호를 검출하는 단계; 및획득된 상기 각 치과 단면 이미지 내의 각 객체의 경계 및 상기 각 치아의 치아 번호를 기초로, 치아별 3차원 이미지를 생성하는 단계를 포함하고,상기 치아별 3차원 이미지는 치아번호에 따라 표시방법이 상이한 3차원 치아 이미지 표시 방법.
- 입력된 3차원 치과 이미지에 대해 액시얼(Axial) 방향으로 분할하여 생성된 치과 단면 이미지 내에서 각 객체의 경계를 분할하도록, 치아 분할 신경망을 학습시키는 단계;상기 생성된 치과 단면 이미지 내의 객체중에서 치아 영역을 인식하고, 상기 치아 영역에 속하는 각 치아의 치아 번호를 검출하도록, 치아 검출 신경망을 학습시키는 단계; 및상기 치아 분할 신경망 및 상기 치아 검출 신경망을 통해 획득된 각 치과 단면 이미지 내의 각 객체의 경계 및 각 치아의 치아 번호를 기초로, 치아별 3차원 이미지를 생성하도록 치아 이미지 획득 신경망을 학습시키는 단계를 포함하는 3차원 치아 이미지 표시 장치 학습 방법.
- 제10 항에 있어서,상기 학습용 치과 단면 이미지는 학습용 3차원 이미지가 일축을 따라 분할된 이미지인 3차원 치아 이미지 표시 장치 학습 방법.
- 제10 항에 있어서,상기 치아 이미지 획득 신경망을 학습시키는 단계는,상기 각 학습용 치과 단면 이미지 내의 각 치아의 경계 정보 및 각 치아별 치아번호를 기초로, 상기 3차원 이미지 내의 동일한 치아번호를 갖는 각 치아별로 각각의 3차원 이미지를 생성하도록 학습시키는 단계를 포함하는 3차원 치아 이미지 표시 장치 학습 방법.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP22849825.9A EP4378416A1 (en) | 2021-07-27 | 2022-07-25 | Three-dimensional tooth image display apparatus and method, and method for training three-dimensional tooth image display apparatus |
CN202280042318.9A CN117615733A (zh) | 2021-07-27 | 2022-07-25 | 三维牙齿图像显示装置和方法以及用于训练三维牙齿图像显示装置的方法 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR1020210098582A KR102607886B1 (ko) | 2021-07-27 | 2021-07-27 | 3차원 치아 이미지 표시 장치 및 방법, 그리고 3차원 치아 이미지 표시 장치를 학습시키는 방법 |
KR10-2021-0098582 | 2021-07-27 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023008849A1 true WO2023008849A1 (ko) | 2023-02-02 |
Family
ID=85087098
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/KR2022/010849 WO2023008849A1 (ko) | 2021-07-27 | 2022-07-25 | 3차원 치아 이미지 표시 장치 및 방법, 그리고 3차원 치아 이미지 표시 장치를 학습시키는 방법 |
Country Status (4)
Country | Link |
---|---|
EP (1) | EP4378416A1 (ko) |
KR (2) | KR102607886B1 (ko) |
CN (1) | CN117615733A (ko) |
WO (1) | WO2023008849A1 (ko) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR102651466B1 (ko) * | 2023-03-20 | 2024-03-27 | 고려대학교 산학협력단 | 내포 함수 신경망을 이용한 파노라마 방사선 사진으로부터의 3차원 치아 재구성 장치 및 방법 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101739091B1 (ko) * | 2014-11-19 | 2017-05-24 | 성균관대학교산학협력단 | 치과용 ct 이미지들을 이용한 3차원 이미지의 재구성 방법 및 장치 |
US20190026599A1 (en) * | 2017-07-21 | 2019-01-24 | Dental Monitoring | Method for analyzing an image of a dental arch |
JP2020516335A (ja) * | 2017-03-17 | 2020-06-11 | トロフィー | 動的歯列弓マップ |
KR20200129509A (ko) * | 2019-05-09 | 2020-11-18 | 오스템임플란트 주식회사 | 치아 부가정보 제공 방법 및 그 장치 |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120258416A1 (en) * | 2011-04-08 | 2012-10-11 | Hongsheng Tong | Method to define, measure, and display mesiodistal angulation and faciolingual inclination of each whole tooth |
KR101903424B1 (ko) | 2017-01-10 | 2018-11-13 | 한국광기술원 | 광단층영상시스템 기반 3d 구강 스캐너 및 이를 이용한 치아 상태 진단 방법 |
KR101839789B1 (ko) * | 2017-08-01 | 2018-03-19 | 주식회사 뷰노 | 치과 영상의 판독 데이터 생성 시스템 |
-
2021
- 2021-07-27 KR KR1020210098582A patent/KR102607886B1/ko active Application Filing
-
2022
- 2022-07-25 CN CN202280042318.9A patent/CN117615733A/zh active Pending
- 2022-07-25 EP EP22849825.9A patent/EP4378416A1/en active Pending
- 2022-07-25 WO PCT/KR2022/010849 patent/WO2023008849A1/ko active Application Filing
-
2023
- 2023-11-23 KR KR1020230164931A patent/KR20230164633A/ko active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101739091B1 (ko) * | 2014-11-19 | 2017-05-24 | 성균관대학교산학협력단 | 치과용 ct 이미지들을 이용한 3차원 이미지의 재구성 방법 및 장치 |
JP2020516335A (ja) * | 2017-03-17 | 2020-06-11 | トロフィー | 動的歯列弓マップ |
US20190026599A1 (en) * | 2017-07-21 | 2019-01-24 | Dental Monitoring | Method for analyzing an image of a dental arch |
KR20200129509A (ko) * | 2019-05-09 | 2020-11-18 | 오스템임플란트 주식회사 | 치아 부가정보 제공 방법 및 그 장치 |
Non-Patent Citations (1)
Title |
---|
HEO, HOON, WON-JUN CHOI, OK-SAM CHAE: "3D Reconstruction System of Teeth for Dental Simulation. ", KIPS TRANSACTIONS: PART B, KOREA INFORMATION PROCESSING SOCIETY, KR, vol. 11B, no. 2, 1 April 2004 (2004-04-01), KR , pages 133 - 140, XP093030485, ISSN: 1598-284X * |
Also Published As
Publication number | Publication date |
---|---|
KR20230016952A (ko) | 2023-02-03 |
CN117615733A (zh) | 2024-02-27 |
EP4378416A1 (en) | 2024-06-05 |
KR102607886B1 (ko) | 2023-11-29 |
KR20230164633A (ko) | 2023-12-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2016108453A1 (ko) | 치아 영상 자동 정합 방법, 이를 위한 장치 및 기록 매체 | |
WO2020040349A1 (ko) | 교정 진단을 위한 치과 영상 분석 방법 및 이를 이용한 장치 | |
WO2022164126A1 (ko) | 구강 스캔 데이터의 크라운 분할을 이용한 구강 스캔 데이터와 컴퓨터 단층촬영 이미지 자동 정합 장치 및 방법 | |
WO2020122357A1 (ko) | 의료영상 재구성 방법 및 그 장치 | |
WO2023008849A1 (ko) | 3차원 치아 이미지 표시 장치 및 방법, 그리고 3차원 치아 이미지 표시 장치를 학습시키는 방법 | |
WO2020226473A1 (ko) | 치아 부가정보 제공 방법 및 그 장치 | |
WO2019231104A1 (ko) | 심층 신경망을 이용하여 영상을 분류하는 방법 및 이를 이용한 장치 | |
WO2017039220A1 (ko) | 치아 교정 계획을 위한 이미지 처리 방법, 이를 위한 장치 및 기록 매체 | |
WO2021034138A1 (ko) | 치매 평가 방법 및 이를 이용한 장치 | |
Chen et al. | Detection of proximal caries lesions on bitewing radiographs using deep learning method | |
WO2022108082A1 (ko) | Ct 영상에서의 치아 분할 시스템 및 방법 | |
WO2021054700A1 (ko) | 치아 병변 정보 제공 방법 및 이를 이용한 장치 | |
WO2020209496A1 (ko) | 치아 오브젝트 검출 방법 및 치아 오브젝트를 이용한 영상 정합 방법 및 장치 | |
WO2023182702A1 (ko) | 디지털 병리이미지의 인공지능 진단 데이터 처리 장치 및 그 방법 | |
WO2023121051A1 (ko) | 환자 정보 제공 방법, 환자 정보 제공 장치, 및 컴퓨터 판독 가능한 기록 매체 | |
WO2009091202A2 (ko) | 트렁케이션 아티팩트를 보정하는 방법 | |
WO2021149918A1 (ko) | 골 연령 추정 방법 및 장치 | |
WO2022055158A1 (ko) | 치아 영상 부분 변환 방법 및 이를 위한 장치 | |
WO2022149664A1 (ko) | 의료 영상 분석 방법 및 이를 이용한 장치 | |
WO2023287041A1 (ko) | 상악동 영상 제공 장치와 방법 및 그 학습 방법 | |
WO2011007998A2 (ko) | 3차원 치열영상 획득방법 | |
WO2023229152A1 (ko) | 인공지능을 적용한 3차원 얼굴스캔 자동매칭장치 및 그 장치의 구동방법, 그리고 매체에 저장된 컴퓨터프로그램 | |
WO2020101265A1 (ko) | 심근 이미지 분석 방법 및 장치 | |
WO2023239077A1 (ko) | 신경관 표시 방법, 컴퓨팅 장치 및 이를 위한 컴퓨터 판독 가능한 기록 매체 | |
WO2024071571A1 (ko) | 3차원 구강 모델의 세그멘테이션 방법 및 그 시스템 |
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: 22849825 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18565345 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202280042318.9 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2022849825 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2022849825 Country of ref document: EP Effective date: 20240227 |