Dental jaw model preparation system and method, shell-shaped dental appliance preparation method

Abstract

The present invention provides a dental model preparation system and method, and a shell-shaped dental appliance preparation method. The dental model preparation system includes: a data acquisition module, which is used to obtain digital three-dimensional dental model data information and the coding information of the hollow information identification code corresponding to the digital three-dimensional dental model; a composite module, which is used to composite the digital three-dimensional dental model data information with the coding information of the hollow information identification code to generate digital three-dimensional dental model composite data information; a preparation module, which is used to prepare a dental model with a hollow information identification code by additive manufacturing according to the digital three-dimensional dental model composite data information. The hollow design of the hollow information identification code on the dental model essentially enhances the contrast between the identification code and other parts during the identification process, improves the identification stability and reliability, reduces the problems of unrecognition and identification errors caused by low contrast in batch identification, and improves the yield and efficiency of the entire production process.

Description

Technical Field

[0001] The present invention belongs to the technical field of dental orthodontics, and more specifically relates to the manufacturing technology of shell-shaped dental orthodontic appliances, and in particular to a preparation system and method of a dental jaw model, a method for identifying coded information of a dental jaw model, and a method for preparing a shell-shaped dental orthodontic appliance. Background Art

[0002] At present, shell-shaped dental appliances are generally produced by hot pressing film forming technology. The whole production process includes the steps of making dental models, pressing films, cutting, prying molds, and post-processing. The process of information transmission in the whole production process is generally as follows: first, the dental model is encoded, and then the code on the dental model is read, and the shell-shaped dental appliance is processed and manufactured according to the coded information on the dental model. The information of the whole production process of the shell-shaped dental appliance depends on the input information when the dental model is made. Therefore, the accuracy of the coded information on the dental model and the accuracy of the recognition are extremely important for the subsequent processes.

[0003] The coded information on the dental model is generally produced by 3D printing, that is, a specific code, such as a QR code, is raised on the base surface of the dental model, and then recognized by a scanner or camera.

[0004] Regarding the encoding form of the above-mentioned dental jaw model, the following problems often exist:

[0005] Because it is additive manufacturing, the material of the raised pattern and the material of the underlying base surface are the same, and the color is basically the same, making the subsequent recognition of the code very difficult. For example, it is necessary to add a backlight or side light source to make the pattern and the base surface have a difference in light and dark to help with recognition. Even so, the stability and reliability of recognition are relatively low, and there is a great risk in applying it to batch rapid recognition, which reduces the production efficiency of the entire production line. The information on the tooth model directly printed by 3D is affected by the printing accuracy and the soft material, and it is easy to have unclear and incomplete characters. For this problem, the accuracy of using traditional morphological visual recognition is low.

[0006] Aiming at the problems existing in the current encoding form of dental models, the present invention proposes corresponding solutions. Summary of the invention

[0007] The main purpose of the present invention is to overcome the defects of the prior art and provide "a system and method for preparing a dental model, a method for identifying the coded information of a dental model, and a method for preparing a shell-shaped dental appliance", which solve the problems of low recognition rate of dental model codes, high risk of rapid batch recognition, and low production efficiency.

[0008] The technical solution provided by the present invention is as follows:

[0009] The present invention provides a dental model preparation system, comprising:

[0010] A data acquisition module, used to acquire data information of a digitized three-dimensional dental model and coding information of a hollow information identification code corresponding to the digitized three-dimensional dental model;

[0011] A compound module, used for compounding the digital three-dimensional dental model data information with the coding information of the hollow information identification code to generate digital three-dimensional dental model compound data information;

[0012] The preparation module is used to prepare a dental model with a hollow information identification code by using additive manufacturing according to the composite data information of the digital three-dimensional dental model.

[0013] Further preferably, the dental model with the hollowed-out information identification code comprises a dental model body, an information portion is arranged on the dental model body, and the information portion is arranged on the hollowed-out information identification code.

[0014] Further preferably, the hollow information identification code includes a plurality of hollow units, wherein the distance between two adjacent hollow units is set to 0.2 mm-5 mm.

[0015] Further preferably, the information portion includes a plurality of non-hollow units, and the height difference between the hollow units and the non-hollow units is set to 0.2 mm-5 mm.

[0016] Further preferably, the distance between the edge contour of the hollow unit close to the outer edge of the information portion and the outer edge of the information portion is set to 0.5mm-10mm.

[0017] Further preferably, the hollow unit is configured so that the grayscale level of the hollow unit in the visual recognition image is within the range of 0-125.

[0018] Further preferably, the information portion is provided with an identification and positioning portion for guiding the identification of the hollow information identification code.

[0019] Further preferably, the identification and positioning part is also used to identify the front and back sides of the dental model body.

[0020] Further preferably, the coded information of the hollow information identification code includes any one or any combination of the following: letters, numbers, graphics.

[0021] More preferably, it further comprises a verification module, and the verification module is used to verify the dental model with the hollow information identification code.

[0022] Further preferably, the verification module includes: a visual recognition unit, used to capture the hollow information identification code on the dental model; a data processing unit, used to process and identify the hollow information identification code captured by the visual recognition unit, and obtain the coding information of the hollow information identification code on the dental model; and a verification unit, used to verify the dental model with the hollow information identification code according to the coding information.

[0023] Further preferably, the data processing unit performs specific operations including: identifying the area where the coding information of the hollow information identification code on the dental model is located; segmenting each character in the area where the coding information is located, and extracting each segmented character; identifying each extracted character to obtain the coding information on the dental model.

[0024] Further preferably, the data processing unit adopts an MLP neural network to recognize each character, specifically including the steps of: inputting the feature vector of the extracted character into a trained MLP neural network; outputting the confidence value of the character corresponding to each character in a character set, the character set being a set of characters used to express the encoded information, and the sum of the confidence values ​​of all characters corresponding to the character in the character set is 1; selecting the character value with the largest confidence value in the character set as the character value of the character.

[0025] Further preferably, the verification unit performs specific operations including: reorganizing or assigning characters in the coded information according to a predetermined verification rule to obtain verification data information; verifying the verification data information using CRC or parity check, and judging whether the dental model with the hollow information identification code is qualified based on the verification result; or, performing a hash operation on the verification data information to obtain a verification hash value; obtaining a reference hash value corresponding to the coded information of the hollow information identification code corresponding to the digitized three-dimensional dental model; comparing the verification hash value with the reference hash value, and judging whether the dental model with the hollow information identification code is qualified based on the comparison result.

[0026] The present invention also provides a method for preparing a dental model, comprising the steps of:

[0027] Acquire data information of a digitized three-dimensional dental model and coding information of a hollowing information identification code corresponding to the digitized three-dimensional dental model;

[0028] Combining the data information of the digitized three-dimensional dental model with the coding information of the hollow information identification code to generate composite data information of the digitized three-dimensional dental model;

[0029] According to the composite data information of the digitized three-dimensional dental model, an additive manufacturing method is used to generate a dental model with a hollow information identification code.

[0030] The present invention also provides a method for identifying coded information of a dental model, comprising the steps of:

[0031] Identify the area where the coded information of the hollow information identification code on the dental model is located;

[0032] Segmenting each character in the area where the coded information is located, and extracting each segmented character;

[0033] Recognize each extracted character to obtain the coding information on the dental model.

[0034] Further preferably, the area where the coded information of the hollow information identification code on the dental model is located specifically includes the steps of: obtaining a dental model image; traversing the dental model image according to a set traversal rule to search for the identification and positioning part on the dental model; obtaining the position information of the identification and positioning part in the dental model image; and calculating the area where the coded information is located based on the relative position relationship between the identification and positioning part and the hollow information identification code.

[0035] Further preferably, the segmentation of each character in the area where the coding information is located specifically includes the steps of: identifying a calibration point, the calibration point being a point pre-set according to a fixed order of each character in the coding information, and the calibration point being pre-set with segmentation feature information relative to each character; and segmenting each character based on the segmentation feature information between the character and the calibration point.

[0036] Further preferably, the extraction of the segmented characters specifically includes the steps of: binarizing each character area to obtain a binary information code area of ​​each character area; repairing the binary information code area by mask expansion; filling the repaired binary information code area with a seed filling algorithm; extracting connected areas of the filled binary information code area; filtering the extracted connected areas according to the area size of the connected areas, the filtered connected areas being the image features of the corresponding characters.

[0037] Further preferably, each character is recognized based on an MLP neural network.

[0038] Further preferably, the MLP neural network recognizes each character, specifically comprising the steps of: inputting the feature vector of the extracted character into the trained MLP neural network; outputting the confidence value of the character corresponding to each character in a character set, the character set being a set of characters used to express the encoded information, the sum of the confidence values ​​of all characters corresponding to the character in the character set being 1; selecting the character value with the largest confidence value in the character set as the character value of the character.

[0039] Further preferably, the specific training steps of the trained MLP neural network include: setting a feature vector of a character input to the MLP neural network, an expected output value corresponding to the character, and an expected error between the output value of the MLP neural network and the expected output value; inputting the feature vector of the character into the MLP neural network for calculation to obtain the output value of the MLP neural network; calculating the error between the output value and the expected output value; determining whether the error is greater than the expected error, and if so, using a back propagation algorithm to iteratively update the parameters of each layer of neurons from the output layer to the input layer in the MLP neural network; and cyclically training the updated MLP neural network until the error is less than the expected error and the training of the MLP neural network is completed.

[0040] Further preferably, the method further includes the step of verifying the identified coded information.

[0041] Further preferably, the verification of the identified coded information specifically includes the steps of: reorganizing or assigning characters in the coded information according to a predetermined verification rule to obtain verification data information; using CRC or parity check to verify the verification recognition result; or, performing a hash operation on the verification data information to obtain a verification hash value; obtaining a reference hash value corresponding to the coded information of the hollow information identification code corresponding to the digitized three-dimensional dental model; comparing the verification hash value with the reference hash value, and verifying the verification recognition result based on the comparison result.

[0042] The present invention also provides a method for preparing a shell-shaped dental appliance, comprising the steps of:

[0043] Making a dental model: using the dental model making system or the dental model making method to make a dental model with a hollow information identification code;

[0044] Manufacturing a shell-shaped dental appliance according to the dental model: identifying the hollow information identification code of the dental model, obtaining production data information for manufacturing the shell-shaped dental appliance, and manufacturing the shell-shaped dental appliance according to the production data information.

[0045] Further preferably, the preparing of the shell-shaped dental appliance based on the production data information specifically includes: selecting a diaphragm for preparing the shell-shaped dental appliance according to the production data information; obtaining hot pressing process parameter information, the hot pressing process parameter information including temperature parameters and pressure parameters; heating the diaphragm according to the temperature parameters, and pressing the heated diaphragm onto the dental model according to the pressure parameters to form a shell-shaped dental appliance to be cut; obtaining cutting process parameter information, retrieving a cutting path, and cutting the shell-shaped dental appliance to be cut according to the cutting path to form a shell-shaped dental appliance.

[0046] Further preferably, the method further comprises: marking the shell-shaped dental appliance according to the production data information.

[0047] The dental and jaw model preparation system and preparation method, the dental and jaw model recognition method, and the shell-shaped dental appliance preparation method provided by the present invention can bring at least one of the following beneficial effects:

[0048] 1. The hollow design of the information identification code on the dental model can essentially enhance the contrast between the identification code and other parts during the identification process, improve the stability and reliability of identification, and reduce the problems of unrecognition and identification errors caused by small contrast differences in batch identification, thereby improving the yield and efficiency of the entire production process.

[0049] 2. The hollow design of the information identification code on the dental model essentially enhances the contrast between the identification code and other parts during the identification process. There is no need to use the background light source in the traditional production process, which saves equipment and energy consumption.

[0050] 3. The hollow structure design of the hollow information identification code on the dental model is used in conjunction with the identification method. The same identification target can be achieved by identifying the mirror image pattern on the reverse side. In the transfer and transportation of materials, there is no need for a specific identifiable surface to face a specific position or a specific direction to obtain the coded information. At the same time, the front and back conditions of the material can also be determined according to the above-mentioned pattern. There is no need for the material to be flipped when reading the code, and the placement position information is also determined for the next process. This saves workstations to a great extent and improves the efficiency of the entire manufacturing process.

[0051] 4. Since the code is a hollow structure, the proportion of the hollow part can be larger, which reduces the adhesion of printing, improves the accuracy of recognition, saves the use of raw materials, and reduces production costs.

[0052] 5. For the recognition of dental models with hollow information identification codes, the present invention provides an information code visual recognition method based on an MLP neural network. This method trains unclear and incomplete information codes in the neural network in advance, and can also perform online learning and training on various abnormal information codes, so as to improve the recognition rate and accuracy of information codes in actual production. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The preferred implementation scheme will be described below in a clear and understandable manner with reference to the accompanying drawings to further illustrate the above-mentioned characteristics, technical features, advantages and their implementation methods.

[0054] Figure 1 A schematic diagram of a dental model preparation system provided by the present invention;

[0055] Figure 2 A schematic diagram of another dental model preparation system provided by the present invention;

[0056] Figure 3 A flow chart of the method for preparing a dental jaw model provided by the present invention;

[0057] Figure 4 A flow chart of another method for preparing a dental model provided by the present invention;

[0058] Figure 5 A schematic diagram of the structure of the dental jaw model provided by the present invention;

[0059] Figure 6 A schematic diagram of another dental model structure provided by the present invention;

[0060] Figure 7 A flow chart of the method for identifying the coded information of the dental model provided by the present invention;

[0061] Figure 8 The MLP neural network training flow chart provided by the present invention;

[0062] Fig. 9 A verification flow chart provided by the present invention;

[0063] Fig.10 This is a flow chart of the method for preparing the shell-shaped dental appliance provided by the present invention. DETAILED DESCRIPTION

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings and other implementation methods can be obtained based on these drawings without creative work.

[0065] The present invention provides a dental model with a hollow information identification code, wherein the coding information of the dental model is designed in a hollow form, thereby improving the recognition rate of the information code of the dental model and thus improving the success rate of preparing a shell-shaped dental appliance. The specific technical solution is shown in the following embodiment.

[0066] Embodiment 1:

[0067] This embodiment provides a dental model preparation system, and its schematic diagram is as follows Figure 1 As shown, it includes a data acquisition module 110 , a compound module 120 and a preparation module 130 .

[0068] The data acquisition module 110 is used to acquire the digital three-dimensional dental model data information and the coding information of the hollow information identification code corresponding to the digital three-dimensional dental model. The digital three-dimensional dental model data information refers to the three-dimensional dental model data information suitable for target correction obtained after the original dental model is corrected, and the coding information of the hollow information identification code refers to the three-dimensional data information after the coding information corresponding to the dental model (for example, the existing coding information represented by a two-dimensional code) is three-dimensionally processed.

[0069] The composite module 120 is used to composite the digital three-dimensional dental model data information with the coding information of the hollow information identification code to generate the composite digital three-dimensional dental model data information. Specifically, the composite module 120 composites the digital three-dimensional dental model data information with the coding information of the three-dimensional hollow information identification code into a three-dimensional model format in a unified data format, for example, STL format, that is, the digital three-dimensional dental model data information and the coding information of the three-dimensional hollow information identification code are combined into a composite 3D data set in a unified data format.

[0070] The preparation module 130 is used to prepare a dental model with a hollow information identification code by means of additive manufacturing according to the composite data information of the digital three-dimensional dental model. Among them, additive manufacturing (AM), commonly known as 3D printing, integrates computer-aided design, material processing and molding technology, and is based on digital model files. It uses software and numerical control systems to stack special metal materials, non-metallic materials and medical biomaterials layer by layer in accordance with extrusion, sintering, melting, light curing, injection, etc. to produce a manufacturing technology for physical objects. Therefore, the preparation module 130 in this embodiment is specifically a 3D printer.

[0071] The prepared dental model with a hollowed-out information identification code comprises a dental model body and an information portion integrally formed with the dental model body, wherein the hollowed-out information identification code is arranged on the information portion.

[0072] Preferably, the hollow information identification code includes a plurality of hollow units, wherein each hollow unit represents the coded information of a hollow information identification code. In order to avoid excessively dense distribution of a plurality of hollow units to affect the recognition accuracy of the coded information, further, the distance between two adjacent hollow units is set to 0.2 mm-5 mm. This distance is only a preferred distance, and those skilled in the art can further transform this distance by optimizing the method of image recognition. For example, compared with this distance, those skilled in the art can reduce the distance between two adjacent hollow units, or increase the distance between two adjacent hollow units.

[0073] The area of ​​the information part other than the area where the hollow information identification code is located is a non-hollow area. Further, the non-hollow area is divided into several non-hollow units by several hollow units. For example, the area between two hollow units is a non-hollow unit. In order to increase the recognition of the hollow unit, further, the height difference between the hollow unit and the non-hollow unit is set to 0.2mm-5mm. For example, the hollow unit is a hollow structure, and the area is surrounded by non-hollow units. The height of the non-hollow unit protrudes upward from the plane where it is located by 0.2mm, or 1mm, or 2.5mm, or 5mm. The setting of its height is mainly aimed at forming a stable structure and a clear contrast with the hollow unit, and under this goal, the height can be appropriately reduced to reduce the cost of 3D printing. For another example, the edge where the non-hollow unit contacts the hollow unit protrudes by 0.2 mm, 1 mm, or 5 mm relative to the horizontal plane where the non-hollow unit is located. For another example, the edge where the non-hollow unit contacts the hollow unit may also be recessed by 0.2 mm relative to the horizontal plane where the non-hollow unit is located. The height difference between the hollow unit and the non-hollow unit may be selected according to actual conditions.

[0074] In other embodiments, the distance between the edge contour of the hollow unit close to the outer edge of the information portion and the outer edge of the information portion is set to 0.5mm-10mm, for example, it can be 0.5mm, it can be 2mm, and it can also be specially adjusted according to the size of the hollow unit. For areas with larger hollow unit areas or areas with finer hollow unit structures, the above distance can be set to 5mm, 8mm or 10mm, etc. With this design structure, the complete presentation of the hollow information identification code on the information portion can be ensured, thereby ensuring the clarity of the image of the hollow information identification code taken, avoiding misreading or misjudgment of the hollow information identification code, and improving the recognition accuracy of the coded information of the hollow information identification code.

[0075] In image recognition, the grayscale level of an image directly affects the recognition accuracy and recognition rate of the image. For this reason, in this embodiment, the hollow unit is further set to have a grayscale level within the range of 0-125 in the visual recognition image. Based on the grayscale level range, the hollow information identification code can be quickly and accurately identified.

[0076] The design of the hollow information identification code is described below. For the specific structure of the dental model with the hollow information identification code, please refer to the description of the following embodiment 3, which will not be repeated in this embodiment.

[0077] The coded information of the hollow information identification code includes any one or any combination of letters, numbers, and graphics. For example, the coded information of the hollow information identification code can be a combination of letters and numbers, a combination of uppercase and lowercase letters, a combination of specific letters and numbers, or just graphics.

[0078] Preferably, the coding information of the hollow information identification code is a combination of uppercase letters and numbers. In order to facilitate the identification of the hollow information identification code, in this embodiment, the following rules are defined for the data format of the coding information of the hollow information identification code:

[0079] In order to ensure relatively stable reading, it is preferred to use 8 characters to represent the coded information. Specifically, in order to eliminate the possibility of misreading, a format of 7 valid bits + 1 check code is used.

[0080] Furthermore, in order to improve the recognition rate of similar numbers and letters, this embodiment provides the following character set, and the characters of the coded information of the hollow information identification code are selected from the character set. The selection rules of the character set are as follows. Considering the extreme case, when numbers and uppercase letters are mixed, for similar numbers and letters, numbers are retained first and letters are removed. Specifically:

[0081]

[0082] According to the above selection rules, the character set consists of 23 characters, and the character set includes [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, E, H, K, M, N, P, T, U, V, W, X, Y].

[0083] The specific structure of the prepared dental model with the hollow information identification code will be described in detail in subsequent embodiments and will not be described in detail in this embodiment.

[0084] The prepared dental model with a hollow information identification code is used to prepare a shell-shaped dental appliance. The quality of the prepared dental model is directly related to the quality of the prepared shell-shaped dental appliance. For example, if the prepared dental model is unqualified, the shell-shaped dental appliance prepared based on it is also unqualified. Therefore, it is necessary to further verify whether the prepared dental model is qualified. If the prepared dental model is found to be unqualified, it is necessary to abandon the prepared dental model and prepare a new dental model.

[0085] Specifically, with respect to the dental model prepared in the present embodiment and provided with a hollow information identification code, if the preparation of the hollow information identification code in the dental model fails (for example, the hollow information identification code is wrong), it also means that the dental model is unqualified. Therefore, in the present embodiment, it is preferred to verify the correctness of the hollow information identification code in the dental model. If it is verified that the hollow information identification code in the dental model is wrong, the dental model still needs to be re-prepared.

[0086] Specifically, the present embodiment further includes a verification module 140, and the verification module 140 is used to verify the dental model with the hollow information identification code.

[0087] like Figure 2 As shown, the verification module 140 includes a visual recognition unit 141 , a data processing unit 142 and a verification unit 143 .

[0088] The visual recognition unit 141 is used to capture the hollow information identification code on the dental model. For example, the visual recognition unit 141 is a camera, which recognizes and obtains the hollow information identification code on the dental model by shooting the dental model image.

[0089] The data processing unit 142 is used to process and identify the hollow information identification code to obtain the coded information on the dental model. Specifically, the data processing unit 142 performs the following operations to obtain the coded information on the dental model.

[0090] Identify the area where the coded information of the hollow information identification code on the dental model is located; segment each character in the area where the coded information is located, and extract each segmented character; identify each extracted character to obtain the coded information on the dental model.

[0091] Preferably, the data processing unit uses an MLP neural network to recognize each character, specifically comprising the steps of:

[0092] 1) Input the feature vector of the extracted character into the trained MLP neural network.

[0093] The MLP neural network consists of three layers: input layer, hidden layer and output layer. The grayscale value of each pixel in the connected area of ​​the segmented extracted characters is normalized as a feature vector, which is passed to the hidden layer. The hidden layer calculates the result through weights and activation function (the activation function is a bitangent function) and sends the result to the output layer.

[0094] 2) Output the confidence value of each character in a character set.

[0095] A character set is a collection of characters used to express coded information, and the sum of the confidence values ​​of all characters corresponding to the characters in the character set is 1.

[0096] The output layer uses the conversion function to calculate the confidence value of each character in the character set. The output layer finally outputs the confidence value of each character in the character set. The conversion function is as follows:

[0097] Conversion function:

[0098] 3) Select the character value with the largest confidence value in the character set as the character value of the character.

[0099] The maximum confidence value in the character set is selected as the actual corresponding character value, and this confidence value is used to confirm whether the output character value is credible. If it is too low, it can be considered that the recognition is wrong or failed.

[0100] For example, taking the recognition of character A as an example to illustrate the recognition process of the MLP neural network, the grayscale value of each pixel in the connected area of ​​character A is normalized and converted into a corresponding feature vector, which is then input into the MLP neural network. After calculation by the MLP neural network, the confidence value of each character in the character set corresponding to the feature vector is output, as shown in the following table:

[0101] 0 1 2 3 4 5 6 7 8 9 A E H K M N P T U V W X Y 0 0 0 0 0.02 0 0 0 0 0 0.91 0.02 0.01 0 0.01 0 0.01 0 0 0.02 0 0 0

[0102] The character value of the character corresponding to the maximum confidence value of 0.91 is selected as the character value of the recognized character to achieve the recognition result.

[0103] The verification unit 143 is used to verify the dental model with the hollow information identification code according to the coding information. This embodiment provides the following two verification methods:

[0104] One verification method is that the verification unit performs the following specific operations for verification: reorganizes or assigns values ​​to the characters in the coded information according to a predetermined verification rule to obtain verification data information; uses CRC or parity check to verify the verification data information, and determines whether the dental model with the hollow information identification code is qualified based on the verification result.

[0105] First, number each character in the character set in the following numerical order:

[0106] 0 1 2 3 4 5 6 7 8 9 A E H K M N P T U V W X Y 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

[0107] Taking the coded information 07H8ANPY as an example, the numbers corresponding to the characters in the valid data bits in the coded information are 0, 7, 12, 8, 10, 15, 16. The sum of the numbers corresponding to the characters in the valid data bits is 68. The remainder 22 of 68 divided by 23 is equal to the number corresponding to the check bit Y. The check is successful, indicating that the hollow information identification code is correct, further proving that the dental model is qualified. On the contrary, if the remainder of the sum of the first seven digits divided by 23 is not equal to the eighth digit, it means that the check fails, indicating that the hollow information identification code is wrong, further proving that the dental model is unqualified.

[0108] Another verification method is that the verification unit performs the following specific operations for verification, see Fig. 9 :

[0109] S910: Obtain verification data information: reorganize or assign values ​​to characters in the coded information according to a predetermined verification rule to obtain verification data information;

[0110] S920: Perform a hash operation on the verification data information to obtain a verification hash value δ1;

[0111] S930: Obtaining a reference hash value δ0 corresponding to the coding information of the hollowing information identification code corresponding to the digital three-dimensional dental model;

[0112] S940: Compare the verification hash value with the reference hash value, and determine whether the dental model with the hollow information identification code is qualified based on the comparison result.

[0113] S950: Identify qualified and enter the next process;

[0114] S960: If the verification hash value is not equal to the reference hash value, an alarm is issued, and the recognition cycle is entered again, or manual intervention is performed. At this time, after re-recognition, the corresponding verification hash value is calculated and compared with the reference hash value. If they are equal, the next process is entered. If not, the dental model is determined to be an unqualified product. For details about the hash verification process, please refer to the specific description of the subsequent embodiments, which is incorporated into this embodiment as an implementation method.

[0115] The dental model preparation system provided in this embodiment has the following beneficial effects:

[0116] 1. The hollow design of the information identification code on the dental model can essentially enhance the contrast between the identification code and other parts during the identification process, improve the stability and reliability of identification, and reduce the problems of unrecognition and identification errors caused by small contrast differences in batch identification, thereby improving the yield and efficiency of the entire production process.

[0117] 2. The hollow design of the information identification code on the dental model essentially enhances the contrast between the identification code and other parts during the identification process. There is no need to use the background light source in the traditional production process, which saves equipment and energy consumption.

[0118] 3. The hollow structure design of the hollow information identification code on the dental model is used in conjunction with the identification method. The same identification target can be achieved by identifying the mirror image pattern on the reverse side. In the transfer and transportation of materials, there is no need for a specific identifiable surface to face a specific position or a specific direction to obtain the coded information. At the same time, the front and back conditions of the material can also be determined according to the above-mentioned pattern. There is no need for the material to be flipped when reading the code, and the placement position information is also determined for the next process. This saves workstations to a great extent and improves the efficiency of the entire manufacturing process.

[0119] 4. Since the code is a hollow structure, the proportion of the hollow part can be larger, which reduces the adhesion of printing, improves the accuracy of recognition, saves the use of raw materials, and reduces production costs.

[0120] 5. The dental model prepared by the dental model preparation system is provided with a hollow information identification code, and the hollow information identification code structure will not be damaged or lost during cleaning.

[0121] 6. Regarding the selection of characters in the coded information of the hollow information identification code, the character set of numbers and uppercase letters provided in this embodiment can not only meet the production needs, but also eliminate the extreme cases of similarity between numbers and uppercase letters, further improving the recognition rate of the dental model information code.

[0122] 7. The information code visual recognition method based on the MLP neural network of the present invention first recognizes the hollow information recognition code of the dental model, and then verifies the dental model according to the recognized coding information. The MLP neural network trains unclear and incomplete information codes in the neural network in advance, and can perform online learning and training on various abnormal information codes, so as to improve the recognition rate and correctness of the information code in actual production, and further improve the accuracy of the dental model verification.

[0123] Embodiment 2:

[0124] Based on the first embodiment, this embodiment provides a method for preparing a dental model, and its flow chart is as follows: Figure 3 As shown, the specific steps include the following steps.

[0125] S310: Acquire data information of the digitized three-dimensional dental model and three-dimensional coding information corresponding to the digitized three-dimensional dental model.

[0126] Among them, the digital three-dimensional dental model data information refers to the three-dimensional dental model data information suitable for target correction obtained after the original dental model has been corrected, and the three-dimensional coding information refers to the three-dimensional data information after three-dimensional processing of the coding information corresponding to the dental model (for example, the existing coding information represented by a two-dimensional code).

[0127] S320: Compounding the data information of the digitized three-dimensional dental model with the three-dimensional coding information to generate compound data information of the digitized three-dimensional dental model.

[0128] Specifically, the digital three-dimensional dental model data information and the three-dimensional coding information are combined into a three-dimensional model format under a unified data format, for example, the STL format, that is, the digital three-dimensional dental model data information and the three-dimensional coding information are combined into a composite 3D data set under a unified data format.

[0129] S330: Based on the composite data information of the digitized three-dimensional dental model, an additive manufacturing method is used to generate a dental model with a hollow information identification code.

[0130] Additive Manufacturing (AM), commonly known as 3D printing, integrates computer-aided design, material processing and molding technology, and is based on digital model files. It uses software and CNC systems to stack special metal materials, non-metallic materials and medical biomaterials layer by layer through extrusion, sintering, melting, photocuring, spraying, etc. to produce physical objects.

[0131] Specifically in this embodiment, the composite 3D data set provided in step S320 is prepared layer by layer, and finally a dental model with a hollow information identification code is prepared.

[0132] The prepared dental model with hollow information identification code includes a dental model body and an information part formed integrally with the dental model body, and the information part is provided with a hollow information identification code. The coded information of the hollow information identification code includes any one or any combination of letters, numbers and graphics.

[0133] For example, the coded information of the hollow information identification code can be a combination of letters and numbers, a combination of uppercase and lowercase letters, a combination of specific letters and numbers, or just graphics.

[0134] Preferably, the coding information of the hollow information identification code is a combination of uppercase letters and numbers. In order to facilitate the identification of the hollow information identification code, in this embodiment, the following rules are defined for the data format of the coding information of the hollow information identification code:

[0135] In order to ensure relatively stable reading, it is preferred to use 8 characters to represent the coded information. Specifically, in order to eliminate the possibility of misreading, a format of 7 valid bits + 1 check code is used.

[0136] Furthermore, in order to improve the recognition rate of similar numbers and letters, this embodiment provides the following character set, and the characters of the encoded information of the hollow information identification code are selected from the character set. The selection rules of the character set are as follows:

[0137] Considering the extreme cases, when numbers and uppercase letters are mixed, numbers are kept first and letters are removed for similar numbers and letters. Specifically:

[0138]

[0139] According to the above selection rules, the character set consists of 23 characters, and the character set includes [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, E, H, K, M, N, P, T, U, V, W, X, Y].

[0140] The specific structure of the prepared dental model with the hollow information identification code will be described in detail in subsequent embodiments and will not be described in detail in this embodiment.

[0141] The prepared dental model with a hollow information identification code is used to prepare a shell-shaped dental appliance. The quality of the prepared dental model is directly related to the quality of the prepared shell-shaped dental appliance. For example, if the prepared dental model is unqualified, the shell-shaped dental appliance prepared based on it is also unqualified. Therefore, it is necessary to further verify whether the prepared dental model is qualified. If the prepared dental model is found to be unqualified, it is necessary to abandon the prepared dental model and prepare a new dental model.

[0142] Specifically, with respect to the dental model prepared in the present embodiment and provided with a hollow information identification code, if the preparation of the hollow information identification code in the dental model fails (for example, the hollow information identification code is wrong), it also means that the dental model is unqualified. Therefore, in the present embodiment, it is preferred to verify the correctness of the hollow information identification code in the dental model. If it is verified that the hollow information identification code in the dental model is wrong, the dental model still needs to be re-prepared.

[0143] like Figure 4 As shown, the dental model preparation method provided in this embodiment further includes step S340.

[0144] S340: Verify the dental model with the hollow information identification code.

[0145] Step S340 specifically includes the following steps:

[0146] S341: Obtain the hollow information identification code on the dental model.

[0147] For example, the dental model image is photographed by a camera to obtain the hollow information identification code on the dental model.

[0148] S342: Process and identify the hollow information identification code to obtain the coded information on the dental model.

[0149] Specifically, the area where the coded information of the hollow information identification code on the dental model is located is identified; each character in the area where the coded information is located is segmented, and each segmented character is extracted; each character is identified to obtain the coded information.

[0150] The steps of identifying each character specifically include the following steps:

[0151] 1) Input the feature vector of the extracted character into the trained MLP neural network.

[0152] The MLP neural network consists of three layers: input layer, hidden layer and output layer. The grayscale value of each pixel in the connected area of ​​the segmented extracted characters is normalized as a feature vector, which is passed to the hidden layer. The hidden layer calculates the result through weights and activation function (the activation function is a bitangent function) and sends the result to the output layer.

[0153] 2) Output the confidence value of each character in a character set.

[0154] A character set is a collection of characters used to express coded information, and the sum of the confidence values ​​of all characters corresponding to the characters in the character set is 1.

[0155] The output layer uses the conversion function to calculate the confidence value of each character in the character set. The output layer finally outputs the confidence value of each character in the character set. The conversion function is as follows:

[0156] Conversion function:

[0157] 3) Select the character value with the largest confidence value in the character set as the character value of the character.

[0158] The maximum confidence value in the character set is selected as the actual corresponding character value, and this confidence value is used to confirm whether the output character value is credible. If it is too low, it can be considered that the recognition is wrong or failed.

[0159] For example, taking the recognition of character A as an example to illustrate the recognition process of the MLP neural network, the grayscale value of each pixel in the connected area of ​​character A is normalized and converted into a corresponding feature vector, which is then input into the MLP neural network. After calculation by the MLP neural network, the confidence value of each character in the character set corresponding to the feature vector is output, as shown in the following table:

[0160] 0 1 2 3 4 5 6 7 8 9 A E H K M N P T U V W X Y 0 0 0 0 0.02 0 0 0 0 0 0.91 0.02 0.01 0 0.01 0 0.01 0 0 0.02 0 0 0

[0161] The character value of the character corresponding to the maximum confidence value of 0.91 is selected as the character value of the recognized character to achieve the recognition result.

[0162] S343: Verify the dental model with the hollow information identification code according to the recognized coding information. This embodiment provides the following two verification methods:

[0163] One verification method is: reorganizing or assigning values ​​to characters in the coded information according to a predetermined verification rule to obtain verification data information; verifying the verification data information using CRC or parity check, and judging whether the dental model with the hollow information identification code is qualified based on the verification result.

[0164] First, number each character in the character set in the following numerical order:

[0165] 0 1 2 3 4 5 6 7 8 9 A E H K H N P T U V W X Y 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

[0166] Taking the coded information 07H8ANPY as an example, the numbers corresponding to the characters in the valid data bits in the coded information are 0, 7, 12, 8, 10, 15, 16. The sum of the numbers corresponding to the characters in the valid data bits is 68. The remainder 22 of 68 divided by 23 is equal to the number corresponding to the check bit Y. The check is successful, indicating that the hollow information identification code is correct, further proving that the dental model is qualified. On the contrary, if the remainder of the sum of the first seven digits divided by 23 is not equal to the eighth digit, it means that the check fails, indicating that the hollow information identification code is wrong, further proving that the dental model is unqualified.

[0167] Another way to check is to see Fig. 9 :According to a predetermined verification rule, characters in the coded information are reorganized or assigned values ​​to obtain verification data information; a hash operation is performed on the verification data information to obtain a verification hash value; a reference hash value corresponding to the coded information of the hollow information identification code corresponding to the digital three-dimensional dental model is obtained; the verification hash value is compared with the reference hash value, and whether the dental model with the hollow information identification code is qualified is determined according to the comparison result. The specific process can be referred to in the above embodiment, which will not be repeated here.

[0168] The beneficial effects that can be achieved by the dental model preparation method provided in this embodiment can be referred to the beneficial effects of the corresponding methods provided above, and will not be repeated here.

[0169] Embodiment three:

[0170] Based on Example 1 and Example 2, this embodiment provides a dental model, which can be prepared using the dental model preparation system provided in Example 1, or can be prepared using the dental model preparation method provided in Example 2. Different from the existing dental models, the dental model provided in this embodiment is accompanied by a hollow information identification code.

[0171] like Figure 5 As shown, the dental model of this embodiment includes a dental model body 510, an information part 520 integrally formed with the dental model body 510, and a hollow information identification code 530 is provided on the information part 520. This embodiment mainly describes the dental model with a hollow information identification code from the structural point of view, and the same parts as those in the first embodiment are not repeated in this embodiment. For details, please refer to the first embodiment, for example, the hollow information identification code 530 includes a plurality of hollow units and a plurality of non-hollow units in which the information part 520 is divided by the hollow units, and a specific description of the hollow units and the non-hollow units.

[0172] The information portion 520 may be a convex structure or a square structure (eg, a square structure or a rectangular structure). The convex structure of the information portion 520 is a preferred solution, as this convex structure can save placement space in 3D printing.

[0173] The integrated structure of the information part 520 and the dental model body 510 may refer to an integrated structure in which the information part 520 and the dental model body 510 are made of the same material or different materials and there is no other spacing between them, or it may refer to an integrated structure in which the information part 520 and the dental model body 510 are made of the same material or different materials and need to be connected by other connecting parts. The structure formed between the information part 520 and the dental model body 510 is determined according to the actual application and is not specifically limited in this embodiment.

[0174] In addition, the information part 520 can be located inside the dental model body 510 or outside the dental model body 510. The position of the information part 520 in the dental model body 510 depends on the actual situation and is not limited in this embodiment.

[0175] In addition to being used to design the hollow information identification code 530, the information portion 520 may also be designed with other functional structures. For example, the information portion 520 is provided with an identification positioning portion 540 for guiding the identification of the hollow information identification code 530. Specifically, the identification positioning portion 540 may be irregularly designed so that the identification positioning portion 540 not only has the function of guiding identification, but also has the function of identifying the front and back sides of the dental model body 510.

[0176] like Figure 5 As shown, the identification and positioning portion 540 includes an identification hole and a V-shaped opening. By setting the positions of the identification hole and the V-shaped opening, for example, taking the front side as the reference plane and the center line of the V-shaped opening as the reference line, the identification hole is set below the V-shaped opening and biased to the left of the center line of the V-shaped opening. Through this design, when the dental model body 510 is in the front position, the identification hole is located below the V-shaped opening and on the left side of the center line of the V-shaped opening. When the dental model body 510 is turned upside down to the back side, the identification hole is located above the V-shaped opening. When the dental model body 510 is turned left and right to the back side, the identification hole is located below the V-shaped opening and on the right side of the center line of the V-shaped opening. The hollow identification code mentioned in the present invention can realize the function of front and back identification, and further, it can provide the position information of the object for the next process. For example, based on the relative position relationship between the identification hole and the V-shaped opening, it can be judged whether the dental model body 510 is on the front side or on the back side. When it is on the back side, it is further judged whether it is turned upside down or turned left and right.

[0177] The specific structure of the identification and positioning portion 540 described above is only a specific embodiment, and those skilled in the art can design an identification and positioning portion 540 with other structures and the same function based on this concept.

[0178] For another example, the information unit 520 may also be provided with a mounting unit 550 for mounting the dental model body 510. Figure 6 As shown, the mounting portion 550 can be specifically designed as a mounting hole. The mounting portion 550 can also have a variety of structures. The mounting portion 550 can be matched with the mounting positioning structure of the dental model carrier plate corresponding to the dental model to achieve the retention of the dental model during the transmission process and the lamination process. Figure 6 As shown in the figure, the dental model is fixed to the dental model supporting plate through two positioning holes.

[0179] The hollow information identification code 530 is the carrier of the dental model information code, and the coding information of the hollow information identification code may include any one or any combination of letters, numbers, and graphics. For example, the coding information of the hollow information identification code may be a combination of letters and numbers, a combination of uppercase and lowercase letters, a combination of specific letters and numbers, or only graphics. The data format definition rules of the coding information of the hollow information identification code and the subsequent coding verification method can refer to the corresponding parts of the corresponding system and method embodiments provided above, and will not be repeated here.

[0180] The characters used in the hollow information identification code 530 are selected from the above-mentioned character set, and the selected characters can be arranged horizontally, vertically, or in other orderly arrangements (for example, circular arrangement) on the information portion 520.

[0181] It is recommended that the font of each character should be 4mm or above in height. Bold and other clear fonts can be used. The height of the characters should be consistent. To meet the requirements of visual inspection, the interval between adjacent characters should be greater than 0.5mm. The font can be selected with a floating height of 0.5mm-0.7mm or a hollow font. The distance between the outer edge of the information part 520 and the edge of the outer contour of the character needs to exceed the character by 0.5mm-10mm, and the thickness should be uniform to ensure the integrity of the information code.

[0182] Furthermore, for the coded information area in the information section 520, each area in the coded information area can be defined, for example, the specific meaning represented by the numbers / letters in each area can be defined, such as the appliance serial number, patient number, case type, etc., and the corresponding medical statistical information of the production process (day / month / year) can be completed by retrieving the marker of the corresponding area by computer; the marker can also have production information identification, such as printer number, diaphragm number, laminator number and other mechanical position numbers in the production process, which can assist in completing the corresponding equipment / machine statistical information in the production process, and then assist production personnel in tracking and maintaining the equipment status.

[0183] In addition, the coding information area may also be relatively marked, and each area may correspond to an area of ​​one character or an area composed of multiple characters, which is not specifically limited in this embodiment.

[0184] Embodiment 4:

[0185] In combination with the dental model with hollow information identification code provided in the previous embodiment of the present invention, this embodiment provides a method for identifying the coded information of the dental model, and its flow chart is as follows: Figure 7 As shown, the specific steps include the following steps.

[0186] S710: Identify the area where the coded information of the hollow information identification code on the dental model is located. Specifically, the steps include:

[0187] S711: Acquire a dental model image. For example, the dental model image may be photographed by a camera to acquire the dental model.

[0188] S712: Traverse the dental model image to find the identification and positioning part on the dental model according to the set traversal rules. Figure 8 As shown, for the dental model image obtained in step S711, the dental model image is traversed from the upper left to the lower right in a "bow" shape to find the identification positioning part.

[0189] S713: Obtain the position information of the identification and positioning part in the dental model image. Calculate the matching degree R(x, y) between the pixel value I(x′, y′) on each overlapping sub-image and the pixel value T(x′, y′) on the pre-saved positioning image, and obtain the position of the identification and positioning part on this dental model image according to the highest matching degree.

[0190] The calculation formula of the matching degree is shown in the following formula (1):

[0191]

[0192] S714: Calculate the area where the coded information is located based on the relative position relationship between the identification and positioning part and the hollow information identification code. Since the relative position of the character area and the identification and positioning part area is fixed, the characteristic area where the information code is located can be directly calculated based on the traversed identification and positioning part position.

[0193] S720: Segment each character in the area where the coded information is located, and extract each segmented character. Segmenting each character in the area where the coded information is located specifically includes the following steps:

[0194] S721: Identify a calibration point, which is a point pre-set according to the fixed order of each character in the encoding information, and the calibration point is pre-set with segmentation feature information relative to each character. For example, the segmentation feature information may be a specific distance between each character and the calibration point.

[0195] S722: Segment each character based on the segmentation feature information between the character and the calibration point. In addition to the above segmentation, the image segmentation method can also be used to segment the information code area in the dental model image through the trained segmentation model. Extract each segmented character, specifically including the following steps:

[0196] S723: Binarize each character area to obtain the binary information code area of ​​each character area. Specifically, the character area is binarized according to the threshold parameter T. If the pixel value f(x, y) of the character area is greater than T, the pixel value is the maximum value 255, otherwise it is the minimum value 0. The final binary area g(x, y) is calculated as shown in formula (2) (the purpose of this step is to remove unnecessary interference and noise). Using different threshold parameters can obtain binary information code areas of different area sizes.

[0197]

[0198] S724: Repair the binary information code area by mask expansion. For example, use a mask m*m element with the same width as the information code, traverse each pixel after binarization, perform an "OR" operation with the mask and the binary image, and repair the previously broken part by mask expansion.

[0199] S725: Fill the restored binary information code area using a seed filling algorithm. For example, using a seed filling method, starting from a specified seed, searching in all directions, comparing each pixel one by one, to extract each connected area in the image.

[0200] S726: Extract connected regions from the filled binary information code region.

[0201] S727: Filter the extracted connected regions according to the size of the connected regions, and the filtered connected regions are the image features of the corresponding characters. Each character region is filtered again according to the number of pixels in the connected regions (i.e., the size of the area), and the filtered connected regions can accurately reflect the image features of each character.

[0202] S730: Recognize each extracted character to obtain the coding information on the dental model. In this embodiment, the MLP neural network is used to recognize each character. The MLP neural network recognizes each character in detail including the following steps:

[0203] S731: Input the feature vector of the extracted character into the trained MLP neural network. The MLP neural network includes a three-layer structure of input layer, hidden layer and output layer. The gray value of each pixel in the connected area of ​​the segmented and extracted character is normalized as a feature vector, and the feature vector is passed to the hidden layer. The hidden layer calculates the result through weights and excitation function (the excitation function is a bitangent function), and sends the result to the output layer.

[0204] S732: Output the confidence value of each character in a character set. The character set is a set of characters used to express coded information, and the sum of the confidence values ​​of all characters corresponding to the character in the character set is 1. The output layer uses a conversion function to calculate the confidence value of each character in the character set, and the output layer finally outputs the confidence value of each character in the character set. The conversion function is shown in formula (3):

[0205]

[0206] S733: Select the character value with the largest confidence value in the character set as the character value of the character.

[0207] The maximum confidence value in the character set is selected as the actual corresponding character value, and this confidence value is used to confirm whether the output character value is credible. If it is too low, it can be considered that the recognition is wrong or failed.

[0208] For example, taking the recognition of character A as an example to illustrate the recognition process of the MLP neural network, the grayscale value of each pixel in the connected area of ​​character A is normalized and converted into a corresponding feature vector, which is then input into the MLP neural network. After calculation by the MLP neural network, the confidence value of each character in the character set corresponding to the feature vector is output, as shown in the following table:

[0209] 0 1 2 3 4 5 6 7 8 9 A E H K M N P T U V W X Y 0 0 0 0 0.02 0 0 0 0 0 0.91 0.02 0.01 0 0.01 0 0.01 0 0 0.02 0 0 0

[0210] The character value of the character corresponding to the maximum confidence value of 0.91 is selected as the character value of the recognized character to achieve the recognition result.

[0211] It should be noted that the above-mentioned MLP neural network is a neural network that has been trained in advance. Furthermore, this embodiment also provides a step for training the MLP neural network, which specifically includes the following steps, and its flow chart is as follows: Figure 8 As shown:

[0212] S810: Setting the feature vector of the MLP neural network input character, the expected output value corresponding to the character, and the expected error between the output value of the MLP neural network and the expected output value.

[0213] S820: Input the feature vector of the character into the MLP neural network for calculation to obtain the output value of the MLP neural network.

[0214] S830: Calculate the error between the output value and the expected output value.

[0215] S840: Determine whether the error is greater than the expected error. If so, execute step S850; if not, execute step S860.

[0216] S850: Use the back propagation algorithm to iteratively update the parameters of each layer of neurons from the output layer to the input layer in the MLP neural network, and go to step S820.

[0217] S860: MLP neural network training is completed.

[0218] Furthermore, in order to determine whether the coded information identified by the above steps is correct, this embodiment also includes the step of verifying the identified coded information. This embodiment provides the following two verification methods:

[0219] One way to verify is:

[0220] The characters in the coded information are reorganized or assigned according to a predetermined verification rule to obtain verification data information; the verification data information is verified by using CRC or parity check, and whether the dental model with the hollow information identification code is qualified is determined based on the verification result.

[0221] First, number each character in the character set in the following numerical order:

[0222] 0 1 2 3 4 5 6 7 8 9 A E H K M N P T U V W X Y 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

[0223] Taking the coded information 07H8ANPY as an example, the numbers corresponding to the characters in the valid data bits in the coded information are 0, 7, 12, 8, 10, 15, 16. The sum of the numbers corresponding to the characters in the valid data bits is 68. The remainder 22 of 68 divided by 23 is equal to the number corresponding to the check bit Y. The check is successful, indicating that the recognized coded information is correct. On the contrary, if the remainder of the sum of the first seven digits divided by 23 is not equal to the eighth digit, it means that the check fails and the recognized coded information is wrong.

[0224] For another verification method, see Fig. 9 :

[0225] S910: Obtain verification data information: a marker is preset, and during verification, the marker is first read, characters in the coded information are read according to the marker, and the characters in the coded information are reorganized or assigned values ​​according to a predetermined verification rule to obtain verification data information;

[0226] S920: Perform a hash operation on the verification data information to obtain a verification hash value δ1; obtain a mathematical expression based on the verification data information, and bring the mathematical expression into the hash function for calculation to obtain a value δ1,

[0227] S930: Obtain the reference hash value δ0 corresponding to the coding information of the hollow information identification code corresponding to the digital three-dimensional dental model: Each embodiment can choose to calculate the reference hash value in the front-end database (such as a host computer or an ERP system or an MES system, etc.) or in the verification module according to the actual production situation. When the reference hash value is calculated in the front-end database, this step only needs to call the reference hash value, and no calculation process is required. When the reference hash value is calculated in the verification process, it is necessary to first call the coding information of the hollow information identification code corresponding to the digital three-dimensional dental model, and then perform hash calculation on the coding information to obtain the reference hash value. The present invention does not specifically limit the specific method for obtaining the reference hash value.

[0228] S940: Compare the verification hash value with the reference hash value, and determine whether the dental model with the hollow information identification code is qualified based on the comparison result.

[0229] S950: Identify qualified and enter the next process;

[0230] S960: If the verification hash value is not equal to the reference hash value, an alarm is issued and the recognition cycle is entered again, or manual intervention is performed. At this time, after re-recognition, the corresponding verification hash value is calculated and compared with the original hash value. If they are equal, the next process is entered. If not, the dental model is judged to be a substandard product.

[0231] For another example, multiple groups of markers can be further set for verification and error correction. Specifically, two groups of markers are set as an example for explanation. First, the first group of markers is read, and the characters in the coded information are read according to the first group of markers. Characters are reorganized or assigned according to certain rules to obtain a first mathematical expression. The first mathematical expression is brought into the hash function for calculation to obtain a value δ1. The value δ1 is compared with the reference value δ0. If the two are not equal, the second group of markers is read, and the characters in the coded information are read according to the second group of markers. Characters are reorganized or assigned according to certain rules to obtain a second mathematical expression. The second mathematical expression is brought into the hash function for calculation to obtain a value δ2. The value δ2 is compared with the reference value δ0. If the two are equal, the verification is successful, indicating that the recognized coded information is correct. On the contrary, if the two are not equal, it means that the verification fails and the recognized coded information is wrong.

[0232] The coded information recognition method provided in this embodiment can train unclear and incomplete information codes in a neural network in advance, and can also perform online learning and training on various abnormal information codes, so as to improve the recognition rate and accuracy of information codes in actual production.

[0233] Embodiment five:

[0234] Based on the first to fourth embodiments, this embodiment provides a method for preparing a shell-shaped dental appliance, and its flow chart is as follows: Fig.10 As shown, the specific steps include the following steps.

[0235] S1010: Make a dental model.

[0236] The dental model can be prepared by the dental model preparation system provided in the first embodiment, or by the dental model preparation method provided in the second embodiment. Regardless of which method is used, the final prepared dental model is a dental model with a hollow information identification code.

[0237] Regarding the specific structural design of the dental model with the hollow information identification code, please refer to the dental model provided in the third embodiment, which will not be described in detail in this embodiment.

[0238] S1020: Make a shell-shaped dental appliance based on the dental model.

[0239] Specifically, the hollow information identification code of the dental model is identified, production data information for preparing the shell-shaped dental appliance is obtained, and the shell-shaped dental appliance is prepared based on the production data information.

[0240] The process of identifying the hollow information identification code on the dental model specifically includes the following steps:

[0241] Identify the area where the coded information of the hollow information identification code on the dental model is located;

[0242] Segment each character in the area where the coded information is located, and extract each segmented character;

[0243] The extracted characters are recognized based on the MLP neural network to obtain the coding information on the dental model.

[0244] The above steps only describe the basic recognition process. For the specific process of recognition, please refer to the coding information recognition method provided in Example 4, which will not be described in detail in this example.

[0245] Based on the recognition result of the coded information in the dental model, the shell-shaped dental appliance can be prepared by the following hot pressing film forming process:

[0246] 1) selecting a membrane for preparing a shell-shaped dental appliance according to production data information, wherein the production data information includes information such as the appliance serial number, patient information, case information, printer number, membrane number, and laminator number;

[0247] 2) obtaining hot pressing process parameter information, where the hot pressing process parameter information includes temperature parameters and pressure parameters;

[0248] 3) heating the diaphragm according to the temperature parameter, and pressing the heated diaphragm onto the prepared dental model according to the pressure parameter to form a shell-shaped dental appliance to be cut;

[0249] 4) Obtain cutting process parameter information, retrieve the cutting path, and cut the shell-shaped dental appliance to be cut according to the cutting path to form the shell-shaped dental appliance.

[0250] After the shell-shaped dental appliance is prepared, the process further includes marking the shell-shaped dental appliance according to the production data information, for example, marking the shell-shaped dental appliance according to the appliance serial number.

[0251] The method for preparing a shell-shaped dental appliance provided in the present application has the advantages that the method of information transmission in the entire production process is based on the hollow information identification code on the dental model, and the hollow information identification code itself has the characteristics of low adhesion and will not be damaged during cleaning. These inherent characteristics of the hollow information identification code improve the accuracy of the coded information of the dental model, thereby improving the accuracy of information transmission in the entire production process of preparing the shell-shaped dental appliance, and improving the success rate of preparing the shell-shaped dental appliance.

[0252] It should be noted that the above embodiments can be freely combined as needed. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention.

Claims

1. A dental model preparation system, characterized in that: include: A data acquisition module, used to acquire the data information of the digitized three-dimensional dental model and the coding information of the hollow information identification code corresponding to the digitized three-dimensional dental model, wherein the coding information of the hollow information identification code refers to the three-dimensional data information after the coding information corresponding to the dental model is three-dimensionally processed, and the hollow information identification code includes a plurality of hollow units, each of which represents the coding information of a hollow information identification code; A compound module, used for compounding the digital three-dimensional dental model data information with the coding information of the hollow information identification code to generate digital three-dimensional dental model compound data information; The preparation module is used to prepare a dental model with a hollow information identification code by using additive manufacturing according to the composite data information of the digital three-dimensional dental model.

2. The dental model preparation system according to claim 1, characterized in that: The dental model with a hollowed-out information identification code comprises a dental model body, an information portion provided on the dental model body, and the hollowed-out information identification code provided on the information portion.

3. The dental model preparation system according to claim 2, characterized in that: The hollow information identification code includes a plurality of hollow units, wherein the distance between two adjacent hollow units is set to 0.2 mm-5 mm.

4. The dental model preparation system according to claim 3, characterized in that: The information portion includes a plurality of non-hollow units, and the height difference between the hollow units and the non-hollow units is set to 0.2 mm-5 mm.

5. The dental model preparation system according to claim 3, characterized in that: The distance between the edge contour of the hollow unit close to the outer edge of the information portion and the outer edge of the information portion is set to 0.5mm-10mm.

6. The dental model preparation system according to claim 4 or 5, characterized in that: The hollowed-out unit is configured such that the grayscale level of the hollowed-out unit in the visual recognition image is within the range of 0-125.

7. The dental model preparation system according to claim 2, characterized in that: The information portion is provided with an identification and positioning portion for guiding the identification of the hollow information identification code.

8. The dental model preparation system according to claim 7, characterized in that: The identification and positioning part is also used to identify the front and back sides of the dental model body.

9. The dental model preparation system according to claim 2, characterized in that: The coded information of the hollow information identification code includes any one or any combination of the following: letters, numbers, and graphics.

10. The dental model preparation system according to claim 9, characterized in that: It also includes a verification module, which is used to verify the dental model with the hollow information identification code.

11. The dental model preparation system according to claim 10, characterized in that: The verification module comprises: A visual recognition unit, used to capture the hollow information identification code on the dental model; A data processing unit, used for processing and identifying the hollow information identification code captured by the visual recognition unit, and obtaining the coding information of the hollow information identification code on the dental model; A verification unit is used to verify the dental model with the hollow information identification code according to the coding information.

12. The dental model preparation system according to claim 11, characterized in that: The data processing unit performs specific operations including: Identify the area where the coded information of the hollow information identification code on the dental model is located; Segmenting each character in the area where the coded information is located, and extracting each segmented character; Recognize each extracted character to obtain the coding information on the dental model.

13. The dental model preparation system according to claim 12, characterized in that: The data processing unit uses an MLP neural network to recognize each character, specifically including the steps of: Input the feature vector of the extracted characters into the trained MLP neural network; Outputting the confidence value of each character corresponding to the character in a character set, wherein the character set is a set of characters used to express the encoding information, and the sum of the confidence values ​​of all characters corresponding to the character in the character set is 1; The character value of the character with the largest confidence value in the character set is selected as the character value of the character.

14. The dental model preparation system according to claim 11, characterized in that: The specific operations performed by the verification unit include: Reorganize or assign values ​​to characters in the coded information according to a predetermined verification rule to obtain verification data information; The verification data information is verified by using CRC or parity check, and whether the dental model with the hollow information identification code is qualified is determined according to the verification result; or, Performing a hash operation on the verification data information to obtain a verification hash value; Obtaining a reference hash value corresponding to the coding information of the hollowing information identification code corresponding to the digital three-dimensional dental model; The verification hash value is compared with the reference hash value, and whether the dental model with the hollow information identification code is qualified is determined based on the comparison result.

15. A method for making a dental model using the dental model making system according to any one of claims 1 to 14, characterized in that: Includes steps: Acquire data information of a digitized three-dimensional dental model and coding information of a hollowing information identification code corresponding to the digitized three-dimensional dental model; Combining the data information of the digitized three-dimensional dental model with the coding information of the hollow information identification code to generate composite data information of the digitized three-dimensional dental model; According to the composite data information of the digitized three-dimensional dental model, an additive manufacturing method is used to generate a dental model with a hollow information identification code.

16. A method for identifying coded information of a dental model prepared by the dental model preparation system according to any one of claims 1 to 14, characterized in that: Includes steps: Identify the area where the coded information of the hollow information identification code on the dental model is located; Segmenting each character in the area where the coded information is located, and extracting each segmented character; Recognize each extracted character to obtain the coding information on the dental model.

17. The method for identifying coded information according to claim 16, characterized in that: The method of identifying the area where the coded information of the hollow information identification code on the dental model is located specifically includes the steps of: Obtaining dental model images; Traversing the dental model image according to a set traversal rule to search for an identification and positioning portion on the dental model; Acquiring position information of the identification and positioning part in the dental model image; The area where the coded information is located is calculated according to the relative position relationship between the identification and positioning portion and the hollow information identification code.

18. The method for identifying coded information according to claim 17, characterized in that: The segmentation of each character in the area where the coded information is located specifically comprises the steps of: Identifying a calibration point, wherein the calibration point is a point pre-set according to a fixed order of each character in the coded information, and the calibration point is pre-set with segmentation feature information relative to each character; Each character is segmented based on the segmentation feature information between the character and the calibration point.

19. The method for identifying coded information according to claim 18, characterized in that: The extraction of the segmented characters specifically comprises the steps of: Performing binarization processing on each character region to obtain a binarized information code region of each character region; Repairing the binary information code area by mask expansion; Use a seed filling algorithm to fill the repaired binary information code area; Extracting connected regions from the filled binary information code region; The extracted connected regions are filtered according to the sizes of the connected regions, and the filtered connected regions are the image features of the corresponding characters.

20. The method for identifying coded information according to claim 16, characterized in that: Each character is recognized based on the MLP neural network.

21. The method for identifying coded information according to claim 20, characterized in that: The MLP neural network recognizes each character, specifically including the steps of: Input the feature vector of the extracted characters into the trained MLP neural network; Outputting the confidence value of each character corresponding to the character in a character set, wherein the character set is a set of characters used to express the encoding information, and the sum of the confidence values ​​of all characters corresponding to the character in the character set is 1; The character value of the character with the largest confidence value in the character set is selected as the character value of the character.

22. The method for identifying coded information according to claim 21, characterized in that: The specific training steps of the trained MLP neural network include: Set the feature vector of the MLP neural network input character, the expected output value corresponding to the character, and the expected error between the output value of the MLP neural network and the expected output value; Inputting the feature vector of the character into the MLP neural network for calculation to obtain the output value of the MLP neural network; Calculating an error between the output value and the expected output value; Determine whether the error is greater than the expected error, and if so, use a back propagation algorithm to iteratively update the parameters of each layer of neurons from the output layer to the input layer in the MLP neural network; The MLP neural network training is cyclically performed based on the update until the error is less than the expected error, and the MLP neural network training is completed.

23. The method for identifying coded information according to claim 16, characterized in that: The method also includes the step of verifying the identified coded information.

24. The method for identifying coded information according to claim 23, characterized in that: The verification of the identified coded information specifically comprises the steps of: Reorganize or assign values ​​to characters in the coded information according to a predetermined verification rule to obtain verification data information; Using CRC or parity check to check the identification result; or, Performing a hash operation on the verification data information to obtain a verification hash value; Obtaining a reference hash value corresponding to the coding information of the hollowing information identification code corresponding to the digital three-dimensional dental model; The verification hash value is compared with the reference hash value, and the verification recognition result is verified according to the comparison result.

25. A method for preparing a shell-shaped dental appliance, characterized in that: Includes steps: Making a dental model: using the dental model making system according to any one of claims 1 to 14 or the dental model making method according to claim 15 to make a dental model with a hollowed-out information identification code; Manufacturing a shell-shaped dental appliance according to the dental model: identifying the hollow information identification code of the dental model, obtaining production data information for manufacturing the shell-shaped dental appliance, and manufacturing the shell-shaped dental appliance based on the production data information.

26. The method for preparing a shell-shaped dental appliance according to claim 25, wherein: The step of preparing a shell-shaped dental appliance based on the production data information specifically includes: Selecting a membrane for preparing the shell-shaped dental appliance according to the production data information; Acquiring hot pressing process parameter information, wherein the hot pressing process parameter information includes temperature parameters and pressure parameters; The diaphragm is heated according to the temperature parameter, and the heated diaphragm is pressed onto the dental model according to the pressure parameter to form a shell-shaped dental appliance to be cut; The cutting process parameter information is obtained, a cutting path is retrieved, and the shell-shaped dental appliance to be cut is cut according to the cutting path to form a shell-shaped dental appliance.

27. The method for preparing a shell-shaped dental appliance according to claim 25, characterized in that: Also includes: The shell-shaped dental appliance is marked according to the production data information.

Images