CN117745939A - Acquisition method, device, equipment and storage medium of three-dimensional digital model - Google Patents
Acquisition method, device, equipment and storage medium of three-dimensional digital model Download PDFInfo
- Publication number
- CN117745939A CN117745939A CN202311737403.XA CN202311737403A CN117745939A CN 117745939 A CN117745939 A CN 117745939A CN 202311737403 A CN202311737403 A CN 202311737403A CN 117745939 A CN117745939 A CN 117745939A
- Authority
- CN
- China
- Prior art keywords
- dimensional
- model
- region
- category
- view image
- 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
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000003860 storage Methods 0.000 title claims abstract description 18
- 238000013145 classification model Methods 0.000 claims abstract description 44
- 238000012549 training Methods 0.000 claims description 28
- 238000002360 preparation method Methods 0.000 claims description 11
- 238000005192 partition Methods 0.000 claims description 7
- 238000013138 pruning Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 5
- 239000007943 implant Substances 0.000 claims description 4
- 238000009877 rendering Methods 0.000 claims description 4
- 238000013316 zoning Methods 0.000 claims 1
- 230000008569 process Effects 0.000 description 12
- 238000010586 diagram Methods 0.000 description 11
- 230000000877 morphologic effect Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 239000003550 marker Substances 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000011960 computer-aided design Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 210000000214 mouth Anatomy 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 239000003826 tablet Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 239000011505 plaster Substances 0.000 description 1
- 230000008707 rearrangement Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000009966 trimming Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Abstract
The disclosure relates to a method, a device, equipment and a storage medium for acquiring a three-dimensional digital model. Acquiring a two-dimensional view image of a three-dimensional grid model; classifying the two-dimensional view images by using a region classification model to obtain region categories of the two-dimensional view images; and back projecting the two-dimensional view image containing the region categories onto the three-dimensional grid model, determining a plurality of region categories of the three-dimensional grid model, and obtaining the three-dimensional digitized model partitioned according to the region categories. Therefore, a plurality of region categories are determined from a two-dimensional view of the three-dimensional grid model by using the region classification model, and the region categories of the three-dimensional grid model are determined in a back projection mode, so that the mode is not interfered by factors such as diversification of local regions in the three-dimensional grid model, quality of a grid curved surface and the like, and the three-dimensional grid model is classified fully automatically and with high precision, and a good foundation is provided for subsequent tooth restoration.
Description
Technical Field
The disclosure relates to the technical field of oral cavity digitization, in particular to a method, a device, equipment and a storage medium for acquiring a three-dimensional digitization model.
Background
In a real-time intraoral scanning application scene, different region types (such as normal tooth types, standby tooth types and gum types) need to be identified from a three-dimensional grid model obtained through scanning, so that a user can conveniently acquire detail information (such as a neck line and a rest jaw position) from grid regions in different region types, and the user can assist the user in tooth restoration based on the detail information of the grid regions.
Therefore, a high-precision and automatic three-dimensional digital model acquisition method is provided, which is a technical problem to be solved urgently, so as to provide a good foundation for the subsequent tooth restoration process.
Disclosure of Invention
In order to solve the technical problems, the present disclosure provides a method, an apparatus, a device and a storage medium for acquiring a three-dimensional digital model.
In a first aspect, the present disclosure provides a method for obtaining a three-dimensional digitized model, the method comprising:
acquiring a plurality of two-dimensional view images of a three-dimensional grid model;
classifying the two-dimensional view image by using a region classification model to obtain a region category of the two-dimensional view image;
and back-projecting the two-dimensional view image containing the region category onto the three-dimensional grid model, and determining the grid patch category of the three-dimensional grid model to obtain the three-dimensional digitized model with the grid patch category partition.
In a second aspect, the present disclosure provides an acquisition apparatus for a three-dimensional digitized model, the apparatus comprising:
the acquisition module is used for acquiring a two-dimensional view image of the three-dimensional grid model;
the classification module is used for classifying the two-dimensional view images by using a region classification model to obtain region categories of the two-dimensional view images;
and the determining module is used for back projecting the two-dimensional view image containing the region category onto the three-dimensional grid model, determining the grid patch category of the three-dimensional grid model and obtaining the three-dimensional digital model with the partition according to the grid patch category.
In a third aspect, embodiments of the present disclosure further provide an electronic device, including:
one or more processors;
storage means for storing one or more programs,
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method provided by the first aspect.
In a fourth aspect, embodiments of the present disclosure also provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method provided by the first aspect.
Compared with the prior art, the technical scheme provided by the embodiment of the disclosure has the following advantages:
the embodiment of the disclosure relates to a method, a device, equipment and a storage medium for acquiring a three-dimensional digital model, wherein a plurality of two-dimensional view images of a three-dimensional grid model are acquired; classifying the two-dimensional view images by using a region classification model to obtain region categories of the two-dimensional view images; and back-projecting the two-dimensional view image containing the region categories onto the three-dimensional grid model, and determining the region categories of the three-dimensional grid model to obtain the three-dimensional digital model with the regions partitioned according to the region categories. Therefore, the region classification model is utilized to determine the region category from the two-dimensional view of the three-dimensional grid model, and the region category of the three-dimensional grid model is determined in a back projection mode, so that the method is not interfered by factors such as the diversification of local regions in the three-dimensional grid model, the quality of a grid curved surface and the like, the three-dimensional grid model is classified fully automatically and with high precision, and a good foundation is provided for the subsequent tooth restoration process.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the principles of the disclosure.
In order to more clearly illustrate the embodiments of the present disclosure or the solutions in the prior art, the drawings that are required for the description of the embodiments or the prior art will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.
Fig. 1 is a schematic flow chart of a method for obtaining a three-dimensional digitized model according to an embodiment of the disclosure;
fig. 2a is a schematic diagram of a two-dimensional view image corresponding to a first viewing angle provided in an embodiment of the disclosure;
fig. 2b is a schematic diagram of a two-dimensional view image corresponding to a second viewing angle provided in an embodiment of the disclosure;
fig. 3a is a schematic diagram of a classification result of a two-dimensional view image corresponding to a first viewing angle according to an embodiment of the disclosure;
fig. 3b is a schematic diagram of a classification result of a two-dimensional view image corresponding to a second viewing angle according to an embodiment of the disclosure;
FIG. 4 is a schematic diagram of a back projection result of a three-dimensional mesh model provided by an embodiment of the present disclosure;
fig. 5 is a schematic flow chart of S130 provided in an embodiment of the disclosure;
fig. 6 is a flowchart of a training method of a region classification model according to an embodiment of the disclosure;
fig. 7 is a schematic structural diagram of an apparatus for acquiring a three-dimensional digitized model according to an embodiment of the disclosure;
fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure.
Detailed Description
In order that the above objects, features and advantages of the present disclosure may be more clearly understood, a further description of aspects of the present disclosure will be provided below. It should be noted that, without conflict, the embodiments of the present disclosure and features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure, but the present disclosure may be practiced otherwise than as described herein; it will be apparent that the embodiments in the specification are only some, but not all, embodiments of the disclosure.
In the related art, a manual mode, a semi-automatic mode and a full-automatic mode are adopted to directly classify and process the three-dimensional grid model so as to obtain a three-dimensional digital model with region categories.
When the three-dimensional grid model is classified by using a manual mode, different region categories are directly identified on the three-dimensional grid model manually based on a digital drawing software tool. However, this sort approach requires a lot of time and effort, and requires the support of the associated marking tools and software, requires high learning costs, and is also a not trivial challenge for the marking personnel.
When the three-dimensional grid model is classified by using a semi-automatic mode, a marker selects a plurality of key points on the three-dimensional grid model, and then a specific graph-cut algorithm is used for determining a plurality of region categories of the three-dimensional grid model. However, the classification method needs to have better robustness for the graph cut algorithm, has higher requirements for the marking technology of the marker, and also consumes certain manual interaction and operation events.
When the three-dimensional grid model is classified by using a full-automatic mode, a plurality of region categories of the three-dimensional grid model are automatically determined directly based on differential geometric features (such as curvature, normal vector and the like) of the three-dimensional grid model. However, the classification mode has higher requirements on the shape of the three-dimensional grid model and the quality of the grid curved surface, and flaws such as noise data and surface defects can influence the accuracy of region classification.
Therefore, the method for acquiring the three-dimensional digital model in the prior art has higher requirements on the marker or the three-dimensional grid model, and the classification precision can not meet the requirement of subsequent tooth restoration when classifying the complex three-dimensional grid model.
In order to solve the above problems, embodiments of the present disclosure provide a method, an apparatus, a device, and a storage medium for acquiring a three-dimensional digitized model.
The method for acquiring the three-dimensional digitized model provided in the embodiment of the present disclosure is described below with reference to fig. 1 to 6. In the embodiment of the disclosure, the method for acquiring the three-dimensional digital model may be executed by an electronic device. The electronic device may include devices with communication functions, such as a tablet computer, a desktop computer, a notebook computer, and the like, and may also include devices simulated by a virtual machine or a simulator.
Fig. 1 is a schematic flow chart of a method for obtaining a three-dimensional digitized model according to an embodiment of the disclosure.
As shown in fig. 1, the method for acquiring the three-dimensional digitized model may include the following steps.
S110, acquiring a two-dimensional view image of the three-dimensional grid model.
In this embodiment, in the process of performing the dental restoration design, a designer needs to give a dental restoration scheme adapted to a user based on different region types of the three-dimensional mesh model, so that the acquisition accuracy of the three-dimensional digital model has an important influence on the dental restoration design.
The three-dimensional grid model can be understood as digital impression data of teeth, and compared with a traditional physical model and a plaster impression, the process of acquiring the digital impression data through an intraoral scanner is quicker and more efficient, and the drying and curing process of the three-dimensional grid model is not required to wait.
The two-dimensional view image is obtained by projecting the three-dimensional grid model at different visual angles.
Specific implementations of S110 include, but are not limited to, the following: acquiring a three-dimensional grid model of the scanned teeth; and projecting the three-dimensional grid model at a plurality of view angles to obtain a two-dimensional view image.
Specifically, in the process of tooth scanning by the three-dimensional scanner, the electronic device performs grid reconstruction on three-dimensional scanning data acquired by the three-dimensional scanner in real time to generate a three-dimensional grid model of the scanned teeth, and then the three-dimensional grid model can be projected at a plurality of view angles based on three-dimensional morphological features of each region on the three-dimensional grid model to obtain a two-dimensional view image.
Three-dimensional topographical features may be understood as differential geometric features, including but not limited to features such as curvature, normal vector, and the like.
For ease of understanding, fig. 2a shows a two-dimensional view image corresponding to a first viewing angle, and fig. 2b shows a two-dimensional view image corresponding to a second viewing angle, which two-dimensional view images may contain different three-dimensional topographical features, and the first and second viewing angles are not equal.
When the three-dimensional grid model is projected, only the region containing the three-dimensional morphological features with high complexity can be projected, the corresponding view angle is determined, and the region containing the three-dimensional morphological features with low complexity is not projected, so that a local region image of the three-dimensional grid model is obtained, and classification of the three-dimensional grid model is realized based on the local region image. Thus, the calculation amount is reduced while the classification accuracy is ensured.
S120, classifying the two-dimensional view image by using the region classification model to obtain the region category of the two-dimensional view image.
In this embodiment, the two-dimensional view images are input into the region classification model, so that the region classification model is utilized to determine the categories of the pixels of the two-dimensional view images, so that each two-dimensional view image is divided into different categories, and the region category of each two-dimensional view image is determined.
The region classification model may include, but is not limited to, a model such as a U-Net model, and a region classification model suitable for a different scene is selected according to a three-dimensional grid model under the scene.
The region class of the two-dimensional view image can be understood as the pixel region class. The region categories of the two-dimensional view image may be one or more of the pre-labeled region categories.
In this embodiment, the step of determining the region class of the two-dimensional view image includes, but is not limited to: calculating class probability data of each pixel region in the two-dimensional view image by using a region classification model, wherein each pixel region corresponds to one class of probability data, and the class probability data comprises probability values of a plurality of region classes; and regarding each pixel area in the two-dimensional view image, taking the area category with the highest probability value in the category probability data as the area category of the corresponding pixel area, and obtaining the area category of the two-dimensional view image.
The pixel region refers to a region formed by one pixel or a plurality of pixels in the two-dimensional view image.
Specifically, the region classification model performs feature processing on each pixel region in the two-dimensional view image, determines probability values of each pixel region belonging to one or more region categories in the two-dimensional view image according to feature processing results, obtains category probability data containing probability values of a plurality of region categories, acquires one or more probability values corresponding to each pixel region from the category probability data for each pixel region, and then selects a region category with the largest probability value as the region category of the corresponding pixel region, thereby obtaining the region category of the two-dimensional view image. It should be noted that, the region classification model is generally preset with one or more region classes, and each pixel region in the two-dimensional view image is processed by the region classification model to determine a probability value of each pixel region corresponding to each preset region class.
Exemplary, the preset region categories of the region classification model include a preparation category, a tooth category and a gum category, there are 5 two-dimensional view images, each two-dimensional view image has three pixel regions, and the category probability data of the region classification model for predicting the a pixel region includes: the preparation type probability is 95%, the tooth type probability is 5% and the gum type probability is 0%, and the preparation type with the 95% probability is used as the area type of the a pixel area; similarly, the region classification model predicts class probability data for a b-pixel region comprising: the probability of the category of the preparation body is 0%, the probability of the tooth category is 85%, and the probability of the gum category is 15%, and the tooth category with the probability of 85% is taken as the area category of the b pixel area; similarly, the region classification model predicts the class probability data of the c pixel region including: the class of the preparation is 10%, the class of the tooth is 0%, and the class of the gum is 90%, and the class of the gum with the 90% probability is used as the class of the c pixel area.
For ease of understanding, fig. 3a shows a schematic diagram of a classification result of a two-dimensional view image corresponding to a first viewing angle, fig. 3b shows a schematic diagram of a classification result of a two-dimensional view image corresponding to a second viewing angle, and as shown in fig. 3a and 3b, region categories of the two-dimensional view image may include a preparation category 310, a tooth category 320, and a gum category 330.
And S130, back-projecting a plurality of two-dimensional view images containing the region categories onto the three-dimensional grid model, and determining the region categories of the three-dimensional grid model to obtain the three-dimensional digital model with the regions partitioned according to the region categories.
It can be understood that, because the region class of the two-dimensional view image is obtained by determining the class of the pixel points of the two-dimensional view image, the two-dimensional view image is back projected onto the three-dimensional grid model in a back projection mode, so that the region class on the three-dimensional grid model can be determined, and the three-dimensional digital model with the regions partitioned according to the region class is further obtained.
The region class of the three-dimensional mesh model can be understood as the class of the mesh surface patch. Because noise and shielding exist in the process of determining a plurality of two-dimensional view images, the problem of error segmentation exists in the region category of the three-dimensional grid model, after the two-dimensional view images are back projected, geometric algorithms such as a graph cut algorithm and the like can be utilized to repair edge details so as to improve the acquisition precision of the three-dimensional digital model.
Further, in determining the three-dimensional digital model with the region category, the three-dimensional digital model with each region category can be rendered and displayed.
Specifically, the electronic device can render and display the three-dimensional digital models according to the region types, so that a user can design and manufacture dental crowns, dental jackets and other restorations based on the three-dimensional digital models of the region types conveniently.
Optionally, where the three-dimensional digital model is an intraoral three-dimensional digital model, the region class on the intraoral three-dimensional digital model includes one or more of a tooth class, a gum class, a preparation class, a implant stem class, a abutment class, an inlay class. In other scenes, the region category of the three-dimensional digital model can also comprise the region category corresponding to the scene and the like.
For ease of understanding, fig. 4 shows a schematic rendering effect of the three-dimensional digitized model, where the three-dimensional digitized model in fig. 4 includes a three-dimensional mesh model corresponding to the preparation class 410, a three-dimensional mesh model corresponding to the tooth class 420, and a three-dimensional mesh model corresponding to the gum class 430.
In this embodiment, after obtaining a three-dimensional digitized model with regions partitioned by region type, the user may scan the oral cavity region corresponding to the three-dimensional mesh model of different region type with different point distances, then construct the three-dimensional mesh model based on the three-dimensional data obtained by rescanning, and design and manufacture a dental crown, a dental socket, etc. restoration by using computer-aided design (Computer Aided Design, CAD) and computer-aided manufacturing (Computer Aided Manufacturing, CAM) software, so as to implement dental restoration for the user.
The method for acquiring the three-dimensional digital model acquires a two-dimensional view image of a three-dimensional grid model; classifying the two-dimensional view images by using a region classification model to obtain region categories of the two-dimensional view images; and back-projecting the two-dimensional view image containing the region categories onto the three-dimensional grid model, and determining the region categories of the three-dimensional grid model to obtain the three-dimensional digital model with the regions partitioned according to the region categories. Therefore, the region classification model is utilized to determine the region category from the two-dimensional view of the three-dimensional grid model, and the region category of the three-dimensional grid model is determined in a back projection mode, so that the method is not interfered by factors such as diversification of local regions in the three-dimensional grid model, quality of a grid curved surface and the like, and therefore the three-dimensional grid model is classified in a full-automatic and high-precision mode, and a good basis is provided for a subsequent tooth restoration process.
In another embodiment of the present disclosure, a specific determination manner of the region class of the three-dimensional mesh model is explained in detail.
Fig. 5 shows a flowchart of S130 provided by an embodiment of the present disclosure.
As shown in fig. 5, the method for acquiring the three-dimensional digitized model may include the following steps.
S510, back-projecting the two-dimensional view image containing the region category onto a three-dimensional grid model, and determining grid regions corresponding to the pixel regions on the three-dimensional grid model, wherein each grid region corresponds to a plurality of pixel regions, and each grid region corresponds to a different two-dimensional view image.
The pixel area refers to an area formed by one pixel or a plurality of pixels in the two-dimensional view image, and each pixel area corresponds to one area category.
Wherein the grid area may be constituted by one or more grid patches, and one grid patch corresponds to one pixel or a plurality of pixels.
Specifically, based on the mapping relationship between the two-dimensional view image and the three-dimensional grid model, after the two-dimensional view image including the region class is back projected onto the three-dimensional grid model, one grid region corresponding to the plurality of pixel regions is determined on the three-dimensional grid model.
It can be appreciated that, since the two-dimensional view image is obtained by mapping the three-dimensional grid model at a plurality of viewing angles, a plurality of pixel regions in different two-dimensional view images collectively correspond to one grid region on the three-dimensional grid model.
S520, voting for each grid region on the three-dimensional grid model, and determining the region category with the largest number of votes as the region category of the corresponding grid region to obtain the region category of the three-dimensional grid model.
It can be understood that, since each pixel region in the two-dimensional view image corresponds to one region category, and one or more pixel regions in the two-dimensional view image correspond to one grid region in the three-dimensional grid model, after the two-dimensional view image is back projected, one or more region categories corresponding to each grid region in the three-dimensional grid model are preliminarily determined, and then, by voting, from the one or more region categories corresponding to each preliminarily determined grid region, the region category with the largest number of votes is selected as the region category corresponding to the grid region.
In order to facilitate understanding, 5 two-dimensional view images are back projected onto a three-dimensional grid model, each two-dimensional view image comprises a D pixel area, after the 5 two-dimensional view images are back projected onto the three-dimensional grid model, the D grid area on the three-dimensional grid model corresponds to the D pixel area, wherein the area category of the D pixel area in the 3 two-dimensional view images is a preliminary body category, the area category of the D pixel area in the 2 two-dimensional view images is a tooth category, the area category corresponding to the D grid area can be preliminarily determined to be one of the preliminary body category and the tooth category through back projection, then, as the D grid area belongs to the preliminary body category and is determined by the 3 two-dimensional view images, the D grid area belongs to the tooth category is determined by the 2 two-dimensional view images, the number of tickets of the D grid area belongs to the preliminary body category is the largest, and the area category of the D grid area is determined to be the preliminary body category.
Therefore, since each pixel region in the two-dimensional view image corresponds to one region category, and one or more pixel regions in the two-dimensional view image correspond to one grid region in the three-dimensional grid model, after the two-dimensional view image is subjected to back projection, the region category of each grid region on the three-dimensional grid model is specifically determined in a voting mode, and therefore the region category of the three-dimensional grid model is accurately obtained.
S530, obtaining the three-dimensional digital model with the subareas according to the area categories.
In this embodiment, specific implementation manners of S530 include, but are not limited to, the following manners: and dividing the three-dimensional grid model according to the region category of the three-dimensional grid model to obtain a three-dimensional digital model with regions partitioned according to the region category.
Specifically, each grid region of the three-dimensional digital model is partitioned according to region types, for example, the three-dimensional digital model partitioned according to the region types is obtained by clustering, namely, boundaries among all the partitions are obtained, that is, the boundaries among all the partitions of the three-dimensional digital model partitioned according to the region types are clear, and then cutting can be performed based on the boundaries.
Optionally, the three-dimensional mesh model is partitioned by one or more region categories of tooth category, gum category, preparation category, implant bar category, abutment category, inlay category, etc.
In yet another embodiment of the present disclosure, a training process of the region classification model is explained in detail.
Fig. 6 shows a flowchart of a training method of a region classification model according to an embodiment of the disclosure.
As shown in fig. 6, the training method of the region classification model may include the following steps.
S610, acquiring a plurality of two-dimensional reference views of the reference three-dimensional model, and acquiring a reference category of each two-dimensional reference view.
Wherein, the reference three-dimensional model can be understood as a grid model having similar three-dimensional morphological features as the three-dimensional grid model.
The two-dimensional reference views are obtained by projecting the reference three-dimensional model under a plurality of view angles. As described above, when the reference three-dimensional model is projected, only the region including the three-dimensional morphological feature with high complexity may be projected, and the corresponding viewing angle may be determined, and the region including the three-dimensional morphological feature with low complexity may not be projected, thereby obtaining a local region image of the reference three-dimensional model, so that classification of the reference three-dimensional model is achieved based on the local region image. Therefore, the dependence on global information of the reference three-dimensional model is reduced, model training resources are reduced, and meanwhile, the classification precision of the regional classification model is guaranteed.
The reference category is category labeling information of different areas on each two-dimensional reference view. Alternatively, the reference categories may be annotated in a manner such as numerical, alphabetical, binary, and the like.
S620, fine tuning the pre-training model based on the plurality of two-dimensional reference views and the reference category of each two-dimensional reference view to obtain the region classification model.
In order to enable the model to be suitable for classifying the three-dimensional mesh model, one or more channels in the pre-trained model may be pruned during fine-tuning of the pre-trained model. Specific implementations of S620 include, but are not limited to, the following:
adjusting the channel number of each convolution kernel in the pre-training model according to the current pruning rate, and determining the pre-training model after current adjustment; classifying the plurality of two-dimensional reference views by using the currently adjusted pre-training model to obtain the prediction category of each two-dimensional reference view; calculating the current precision of the pre-training model according to the prediction category and the reference category; if the current precision does not reach the preset precision threshold, continuously adjusting the channel number of each convolution kernel in the pre-training model according to the next pruning rate, and returning to execute the steps, and calculating the next precision of the pre-training model until the next precision reaches the preset precision threshold to obtain the region classification model.
Therefore, when the region classification model is trained, the pretraining network is subjected to iterative fine adjustment according to different pruning rates by utilizing a plurality of two-dimensional reference views and the reference category of each two-dimensional reference view, so that the region classification model can be well applied to classifying the three-dimensional grid model, and a high-precision region segmentation model is obtained, so that the method is applicable to automatically and accurately determining the region category of the three-dimensional grid model.
The embodiment of the disclosure further provides an acquisition device for the three-dimensional digital model, which is used for implementing the acquisition method for the three-dimensional digital model, and is described below with reference to fig. 7. In an embodiment of the present disclosure, the acquiring device of the three-dimensional digitized model may be an electronic device. The electronic device may include devices with communication functions, such as a tablet computer, a desktop computer, a notebook computer, and the like, and may also include devices simulated by a virtual machine or a simulator.
Fig. 7 is a schematic structural diagram of an apparatus for acquiring a three-dimensional digitized model according to an embodiment of the disclosure.
As shown in fig. 7, the acquisition apparatus 700 of the three-dimensional digitized model may include:
an acquisition module 710, configured to acquire a two-dimensional view image of the three-dimensional mesh model;
the classification module 720 is configured to perform classification processing on the two-dimensional view image by using a region classification model, so as to obtain a region class of the two-dimensional view image;
and the determining module 730 is configured to back-project the two-dimensional view image containing the region category onto the three-dimensional grid model, determine the region category of the three-dimensional grid model, and obtain a three-dimensional digitized model with regions partitioned according to the region category.
The device for acquiring the three-dimensional digital model acquires a two-dimensional view image of a three-dimensional grid model; classifying the two-dimensional view images by using a region classification model to obtain region categories of the two-dimensional view images; and back-projecting the two-dimensional view image containing the region categories onto the three-dimensional grid model, and determining the region categories of the three-dimensional grid model to obtain the three-dimensional digital model with the regions partitioned according to the region categories. Therefore, the region classification model is utilized to determine the region category from the two-dimensional view of the three-dimensional grid model, and the region category of the three-dimensional grid model is determined in a back projection mode, so that the method is not interfered by factors such as the diversification of local regions in the three-dimensional grid model, the quality of a grid curved surface and the like, the three-dimensional grid model is classified fully automatically and with high precision, and a good foundation is provided for the subsequent tooth restoration process.
In some embodiments of the present disclosure, the obtaining module 710 includes:
a three-dimensional grid model acquisition unit for acquiring a three-dimensional grid model of the scanned teeth;
and the projection unit is used for projecting the three-dimensional grid model at a plurality of view angles to obtain a plurality of two-dimensional view images.
In some embodiments of the present disclosure, classification module 720 includes:
the computing unit is used for computing the class probability data of each pixel area in the two-dimensional view image by utilizing the area classification model, wherein each pixel area corresponds to one class probability data, and the class probability data comprises probability values of a plurality of area classes;
and the region type determining unit is used for taking the region type with the largest probability value in the type probability data as the region type of the corresponding pixel region for each pixel region in the two-dimensional view image, and obtaining the region type of the two-dimensional view image.
In some embodiments of the present disclosure, the determining module 730 includes:
a grid region determining unit, configured to back-project a two-dimensional view image including the region category onto the three-dimensional grid model, and determine grid regions corresponding to pixel regions on the three-dimensional grid model, where each grid region corresponds to a plurality of pixel regions, and each grid region corresponds to a different two-dimensional view image;
and the voting unit is used for voting the area categories of the pixel areas corresponding to each grid area for each grid area on the three-dimensional grid model, determining the area category with the largest number of votes as the area category of the corresponding grid area, and obtaining the area category of the three-dimensional grid model.
In some embodiments of the present disclosure, the determining module 730 further includes:
the segmentation unit is used for segmenting the three-dimensional grid model according to the region category of the three-dimensional grid model to obtain the three-dimensional digital model with the region division according to the region category.
In some embodiments of the present disclosure, the apparatus further comprises:
and the rendering and displaying module is used for rendering and displaying the three-dimensional digital model.
In some embodiments of the present disclosure, where the three-dimensional digitized model is an intraoral three-dimensional digitized model, the region class on the intraoral three-dimensional digitized model includes one or more of a tooth class, a gum class, a preparation class, a implant stem class, a abutment class, an inlay class.
In some embodiments of the present disclosure, the apparatus further comprises:
the training data acquisition module is used for acquiring a plurality of two-dimensional reference views of the reference three-dimensional model and acquiring reference types of each two-dimensional reference view;
and the fine tuning module is used for fine tuning the pre-training model based on the two-dimensional reference views and the reference category of each two-dimensional reference view to obtain the region classification model.
In some embodiments of the present disclosure, a trimming module includes:
the adjusting unit is used for adjusting the channel number of each convolution kernel in the pre-training model according to the current pruning rate and determining the pre-training model after current adjustment;
the prediction unit is used for classifying the plurality of two-dimensional reference views by using the pre-training model after current adjustment to obtain the prediction category of each two-dimensional reference view;
a calculating unit, configured to calculate a current accuracy of the pre-training model according to the prediction category and the reference category;
and the iteration unit is used for continuously adjusting the channel number of each convolution kernel in the pre-training model according to the next pruning rate if the current precision does not reach the preset precision threshold, and returning to execute the steps, calculating the next precision of the pre-training model until the next precision reaches the preset precision threshold, so as to obtain the region classification model.
It should be noted that, the apparatus 700 for acquiring a three-dimensional digitized model shown in fig. 7 may perform the steps in the method embodiments shown in fig. 1 to 6, and implement the processes and effects in the method embodiments shown in fig. 1 to 6, which are not described herein.
Fig. 8 shows a schematic structural diagram of an electronic device according to an embodiment of the disclosure.
As shown in fig. 8, the electronic device may include a processor 801 and a memory 802 storing computer program instructions.
In particular, the processor 801 may include a Central Processing Unit (CPU), or an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or may be configured to implement one or more integrated circuits of embodiments of the present application.
Memory 802 may include mass storage for information or instructions. By way of example, and not limitation, memory 802 may include a Hard Disk Drive (HDD), floppy Disk Drive, flash memory, optical Disk, magneto-optical Disk, magnetic tape, or universal serial bus (Universal Serial Bus, USB) Drive, or a combination of two or more of these. Memory 802 may include removable or non-removable (or fixed) media, where appropriate. The memory 802 may be internal or external to the integrated gateway device, where appropriate. In a particular embodiment, the memory 802 is a non-volatile solid-state memory. In a particular embodiment, the Memory 802 includes Read-Only Memory (ROM). The ROM may be mask-programmed ROM, programmable ROM (PROM), erasable PROM (Electrical Programmable ROM, EPROM), electrically erasable PROM (Electrically Erasable Programmable ROM, EEPROM), electrically rewritable ROM (Electrically Alterable ROM, EAROM), or flash memory, or a combination of two or more of these, where appropriate.
The processor 801 performs the steps of the method for acquiring a three-dimensional digitized model provided by the embodiments of the present disclosure by reading and executing computer program instructions stored in the memory 802.
In one example, the electronic device may also include a transceiver 803 and a bus 804. As shown in fig. 8, the processor 801, the memory 802, and the transceiver 803 are connected and communicate with each other through a bus 804.
Bus 804 includes hardware, software, or both. By way of example, and not limitation, the buses may include an accelerated graphics port (Accelerated Graphics Port, AGP) or other graphics BUS, an enhanced industry standard architecture (Extended Industry Standard Architecture, EISA) BUS, a Front Side BUS (FSB), a HyperTransport (HT) interconnect, an industry standard architecture (Industrial Standard Architecture, ISA) BUS, an InfiniBand interconnect, a Low Pin Count (LPC) BUS, a memory BUS, a micro channel architecture (Micro Channel Architecture, MCa) BUS, a peripheral control interconnect (Peripheral Component Interconnect, PCI) BUS, a PCI-Express (PCI-X) BUS, a serial advanced technology attachment (Serial Advanced Technology Attachment, SATA) BUS, a video electronics standards association local (Video Electronics Standards Association Local Bus, VLB) BUS, or other suitable BUS, or a combination of two or more of these. Bus 804 may include one or more buses, where appropriate. Although embodiments of the present application describe and illustrate a particular bus, the present application contemplates any suitable bus or interconnect.
The following are embodiments of a computer readable storage medium provided in the embodiments of the present disclosure, where the computer readable storage medium belongs to the same inventive concept as the method for acquiring a three-dimensional digitized model in each of the above embodiments, and details of the method for acquiring a three-dimensional digitized model are not described in detail in the embodiments of the computer readable storage medium, and reference may be made to the embodiments of the method for acquiring a three-dimensional digitized model.
The present embodiment provides a storage medium containing computer executable instructions which, when executed by a computer processor, are used to perform a method of acquiring a three-dimensional digitized model, the method comprising:
acquiring a two-dimensional view image of a three-dimensional grid model;
classifying the two-dimensional view image by using a region classification model to obtain a plurality of region categories of the two-dimensional view image;
and back projecting the two-dimensional view image containing the region category onto the three-dimensional grid model, and determining the region category of the three-dimensional grid model to obtain a three-dimensional digital model with the region category partition.
Of course, the storage medium containing the computer executable instructions provided in the embodiments of the present disclosure is not limited to the above method operations, but may also perform the related operations in the method for obtaining the three-dimensional digitized model provided in any embodiment of the present disclosure.
From the above description of embodiments, it will be apparent to those skilled in the art that the present disclosure may be implemented by means of software and necessary general purpose hardware, but may of course also be implemented by means of hardware, although in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present disclosure may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a FLASH Memory (FLASH), a hard disk, or an optical disk of a computer, where the instructions include a computer cloud platform (which may be a personal computer, a server, or a network cloud platform, etc.) to execute the method for obtaining the three-dimensional digital model provided by the embodiments of the present disclosure.
Note that the above is only a preferred embodiment of the present disclosure and the technical principle applied. Those skilled in the art will appreciate that the present disclosure is not limited to the particular embodiments described herein, and that various obvious changes, rearrangements and substitutions can be made by those skilled in the art without departing from the scope of the disclosure. Therefore, while the present disclosure has been described in connection with the above embodiments, the present disclosure is not limited to the above embodiments, but may include many other equivalent embodiments without departing from the spirit of the present disclosure, the scope of which is determined by the scope of the appended claims.
Claims (12)
1. A method for obtaining a three-dimensional digitized model, comprising:
acquiring a two-dimensional view image of a three-dimensional grid model;
classifying the two-dimensional view image by using a region classification model to obtain a region category of the two-dimensional view image;
and back projecting the two-dimensional view image containing the region category onto the three-dimensional grid model, and determining the region category of the three-dimensional grid model to obtain a three-dimensional digital model with the region category partition.
2. The method of claim 1, wherein the acquiring a two-dimensional view image of a three-dimensional mesh model comprises:
acquiring a three-dimensional grid model of the scanned teeth;
and projecting the three-dimensional grid model at a plurality of view angles to obtain a plurality of two-dimensional view images.
3. The method according to claim 1, wherein classifying the two-dimensional view image by using a region classification model to obtain a region class of the two-dimensional view image comprises:
calculating class probability data of each pixel region in the two-dimensional view image by using the region classification model, wherein each pixel region corresponds to class probability data, and the class probability data comprises probability values of a plurality of region classes;
and regarding each pixel area in the two-dimensional view image, taking the area category with the largest probability value in the category probability data as the area category of the corresponding pixel area, and obtaining the area category of the two-dimensional view image.
4. The method of claim 1, wherein the back projecting the two-dimensional view image containing the region class onto the three-dimensional mesh model, determining the region class of the three-dimensional mesh model, comprises:
back-projecting the two-dimensional view image containing the region category onto the three-dimensional grid model, and determining grid regions corresponding to the pixel regions on the three-dimensional grid model, wherein each grid region corresponds to a plurality of pixel regions, and the plurality of pixel regions corresponding to each grid region correspond to different two-dimensional view images;
voting for each grid region on the three-dimensional grid model, and determining the region category with the highest number of votes as the region category of the corresponding grid region to obtain the region category of the three-dimensional grid model.
5. The method of claim 1, wherein the obtaining a three-dimensional digitized model with zoning by zone category comprises:
and dividing the three-dimensional grid model according to the region category of the three-dimensional grid model to obtain the three-dimensional digital model with the region division according to the region category.
6. The method as recited in claim 1, further comprising:
rendering and displaying the three-dimensional digital model.
7. The method of claim 1, wherein, where the three-dimensional digitized model is an intraoral three-dimensional digitized model, the region class on the intraoral three-dimensional digitized model comprises one or more of a tooth class, a gum class, a preparation class, a implant stem class, a abutment class, an inlay class.
8. The method as recited in claim 1, further comprising:
acquiring a plurality of two-dimensional reference views of a reference three-dimensional model, and acquiring a reference category of each two-dimensional reference view;
and fine tuning the pre-training model based on the plurality of two-dimensional reference views and the reference category of each two-dimensional reference view to obtain the region classification model.
9. The method of claim 8, wherein the fine-tuning the pre-training model based on the plurality of two-dimensional reference views and the reference class of each two-dimensional reference view to obtain the region classification model comprises:
adjusting the channel number of each convolution kernel in the pre-training model according to the current pruning rate, and determining the pre-training model after current adjustment;
classifying the plurality of two-dimensional reference views by using the currently adjusted pre-training model to obtain the prediction category of each two-dimensional reference view;
calculating the current precision of the pre-training model according to the prediction category and the reference category;
if the current precision does not reach the preset precision threshold, continuously adjusting the channel number of each convolution kernel in the pre-training model according to the next pruning rate, and returning to execute the steps, calculating the next precision of the pre-training model until the next precision reaches the preset precision threshold, so as to obtain the region classification model.
10. An acquisition device of a three-dimensional digitized model, characterized by comprising:
the acquisition module is used for acquiring a two-dimensional view image of the three-dimensional grid model;
the classification module is used for classifying the two-dimensional view images by using a region classification model to obtain region categories of the two-dimensional view images;
and the determining module is used for back projecting the two-dimensional view image containing the region category onto the three-dimensional grid model, determining the region category of the three-dimensional grid model and obtaining the three-dimensional digitized model with the region category partition.
11. An electronic device, comprising:
a processor;
a memory for storing executable instructions;
wherein the processor is configured to read the executable instructions from the memory and execute the executable instructions to implement the method of any of the preceding claims 1-9.
12. A computer readable storage medium, on which a computer program is stored, characterized in that the storage medium stores a computer program, which, when executed by a processor, causes the processor to implement the method of any of the preceding claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311737403.XA CN117745939A (en) | 2023-12-15 | 2023-12-15 | Acquisition method, device, equipment and storage medium of three-dimensional digital model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311737403.XA CN117745939A (en) | 2023-12-15 | 2023-12-15 | Acquisition method, device, equipment and storage medium of three-dimensional digital model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117745939A true CN117745939A (en) | 2024-03-22 |
Family
ID=90282441
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311737403.XA Pending CN117745939A (en) | 2023-12-15 | 2023-12-15 | Acquisition method, device, equipment and storage medium of three-dimensional digital model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117745939A (en) |
-
2023
- 2023-12-15 CN CN202311737403.XA patent/CN117745939A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110163198B (en) | Table identification reconstruction method and device and storage medium | |
| CN109409335B (en) | Image processing method, image processing device, computer readable medium and electronic equipment | |
| WO2021238673A1 (en) | Tooth scanning method and apparatus, computer device, and computer-readable storage medium | |
| CA3114650C (en) | Method and apparatus for generating three-dimensional model, device, and storage medium | |
| CN110675487B (en) | Three-dimensional face modeling and recognition method and device based on multi-angle two-dimensional face | |
| US11850113B2 (en) | Systems and methods for constructing a three-dimensional model from two-dimensional images | |
| JP2006506164A5 (en) | ||
| US11250580B2 (en) | Method, system and computer readable storage media for registering intraoral measurements | |
| WO2021189848A1 (en) | Model training method and apparatus, cup-to-disc ratio determination method and apparatus, and device and storage medium | |
| CN116628834A (en) | Contour segmentation correction method and device based on neural network | |
| CN114638838A (en) | Tooth model generation method, device, equipment and storage medium | |
| CN111383207B (en) | Method, system and device for processing musculoskeletal ultrasound image and readable storage medium | |
| CN113780201B (en) | Hand image processing method and device, equipment and medium | |
| CN109492540B (en) | Face exchange method and device in image and electronic equipment | |
| CN114359545A (en) | Image area identification method and device and electronic equipment | |
| CN117745939A (en) | Acquisition method, device, equipment and storage medium of three-dimensional digital model | |
| CN109035380B (en) | Face modification method, device and equipment based on three-dimensional reconstruction and storage medium | |
| CN112270747A (en) | Face recognition method and device and electronic equipment | |
| CN110688704A (en) | Home decoration design method and system and electronic equipment | |
| CN112819741B (en) | Image fusion method and device, electronic equipment and storage medium | |
| CN114187303A (en) | Oral cavity image processing method, device, electronic equipment and computer storage medium | |
| CN109377556B (en) | Face image feature processing method and device | |
| CN117456088A (en) | Method, device, equipment and storage medium for acquiring three-dimensional model | |
| CN116245804A (en) | Foreign matter identification method, device, equipment and storage medium for intraoral scanner | |
| CN116894929A (en) | Method, apparatus and storage medium for reconstructing a complete three-dimensional tooth from a single Zhang Shanmu image |
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 |