JP7466970B1 - Denture size evaluation system and its program - Google Patents

Abstract

[Problem] To provide a denture dimension evaluation system, denture dimension evaluation method, and denture dimension evaluation program that enable denture production by converting qualitative and vague instructions written in a dental technical instruction manual prepared by a dentist into objective and quantitative manufacturing dimensions for dentures. [Solution] A denture dimension evaluation system 1 includes denture dimension evaluation models 25a, 25b trained using teacher data sets 8a-8c in order to convert the denture manufacturing level indicated in the dental technical instruction manual prepared by the dentist into denture manufacturing dimensions, and an evaluation unit 5 that inputs dentition data 15a-15c, contact data 16a-16c, bite data 17a-17c, and fit data 18a-18c extracted by a feature extraction unit 4a to the evaluation model, and outputs the evaluated denture dimensions as output data sets 20a-20c. [Selected Figure] Figure 1

Description

The present invention relates to a denture dimension evaluation system, method, and program that enables denture production by converting the instructions written in a dental laboratory instruction sheet prepared by a dentist into denture manufacturing dimensions.

Conventionally, when treating caries, dental clinics first take impressions of the patient's teeth in order to create dental prostheses and other dental prostheses required for treatment, and then send the impressions along with a document called a dental laboratory instruction sheet to a dental laboratory to request the creation of dentures. The dental laboratory instruction sheet contains instructions from the dentist necessary for the creation of the patient's prosthesis, and the dental laboratory creates a dental model based on the impressions and instructions in the dental laboratory instruction sheet. The dental laboratory then builds a dental prosthesis out of wax on the dental model, takes an impression of the wax prosthesis to create a mold, and then pours metal or other materials into the mold to obtain the dental prosthesis.
However, dental laboratories often have multiple dental clinics as clients, and it is cumbersome to have to manage the dental laboratory instructions sent from the dental clinic for each dentist and patient.In addition, the completed dental prostheses also need to be managed for each dental clinic and patient, which also complicates management.
For example, Patent Document 1 discloses an invention called "Dental technique ordering system, dental technique ordering program, and dental technique ordering method," in which instructions are created and received via the terminals of the dental clinic and dental laboratory, and the instructions include dental technique type information, treatment position information, and material information as instruction information, and the memory unit of the system stores this instruction information in the memory unit DB, and also stores master data on the technical fees corresponding to this information.Then, the calculation unit of the system calculates the technical fees while referring to the instruction information and the master data, and by creating dental technique instructions using a terminal, management is changed from paper to data, and the technical techniques are ordered and received using the data, making it possible to reduce the complexity of management.
Also, Patent Document 2 discloses an instruction management system that electronically manages paper dental laboratory instructions, under the title of an invention "Instruction Management System, Instruction Management Program, and Instruction Management Method." This invention makes it possible to update instruction information written in dental laboratory instructions, and therefore makes it possible to transmit and receive information between dental clinics and dental laboratories via a laboratory terminal and a dental clinic terminal, further reducing the complexity of management.
Furthermore, Patent Document 3 discloses an invention entitled "Technique Identification System, Technique and Program for Identifying Dental Technique." This invention relates to a technique for identifying which patient a completed dental technique belongs to, and specifically, discloses a technique for determining the similarity between a feature based on reference shape data, which is STL data associated with a dental technique instruction sheet, and a feature based on the shape data of a dental technique to be identified. This invention makes it possible to identify which patient's dental technique belongs to by determining the similarity between these two feature amounts.

JP 2023-3999 A Patent No. 6732354 JP 2022-146486 A

However, while the inventions disclosed in these patent documents reduce the complexity of managing dental laboratory instructions and the dental prostheses produced using them, they cannot eliminate the complexity of understanding the instructions written in the dental laboratory instructions prior to management, which reflect the dentist's individual instructions.
For example, in the past, dentists who filled out dental laboratory instructions were often extremely busy, and the instructions were not always written clearly because they had to be completed quickly. Furthermore, the content of the instructions could be ambiguous due to variations in the characteristics of each dentist, so apart from management, there were cases where it was difficult to understand the contents of the dental laboratory instructions in the first place.
Specifically, while the dental laboratory instructions indicate the position of the patient's treatment teeth in the dental arch, they also provide instructions regarding the size of the gap between the treatment tooth and the adjacent tooth, called "contact," instructions regarding the size of the gap between the treatment tooth and the opposing tooth, and instructions regarding the size of the gap between the treatment tooth and the prosthesis, called "fit," such as "tight" and "loose," because the terms expressing characteristics are themselves subjective and qualitative and ambiguous, based on the dentist's own characteristics. This makes it difficult for the dental laboratory to objectively and quantitatively determine the dimensions of the dentures to be made. At the same time, even if a dental laboratory delivers a fabricated product to a dental clinic, it may not match the dentist's intention, and the dental laboratory may have to make corrections or make the product again, which results in complexity and inefficiency in terms of yield.

The present invention was made to address such conventional circumstances, and aims to provide a denture dimension evaluation system, denture dimension evaluation method, and denture dimension evaluation program that enables denture production by converting subjective, qualitative, and vague instructions based on instruction characteristics written in dental laboratory instructions prepared by dentists into objective, quantitative, and clear manufacturing dimensions for dentures.

In order to achieve the above-mentioned object, the first invention is a denture dimension evaluation system for converting the production level for a denture indicated in a dental laboratory instruction sheet created by a dentist into production dimensions, and enabling the production of a denture based on the production dimensions to be reduced in accordance with the characteristics of the instruction from the dentist, and characterized in that it has a denture dimension evaluation model that has been trained using a combination of dentition data of the production level for a learning denture created by the dentist, features formed by adding at least one of three data items of contact data, bite data, and fit data of the production level, and production dimensions that have been assigned in advance corresponding to the features of the learning denture, a feature extraction unit that extracts the feature of the production level for the denture indicated in the dental laboratory instruction sheet, an evaluation unit that evaluates the production dimensions for the denture by applying the feature extracted by the feature extraction unit to the denture dimension evaluation model, and an output unit that outputs the production dimensions for the denture evaluated by the evaluation unit.
In the denture dimension evaluation system configured as above, the feature extraction unit operates to extract features consisting of dentition data of the production level for the denture entered by the dentist in the dental technician instruction sheet and at least one of the three data of contact data, bite data, and fit data of the production level. Also, the evaluation unit operates to input the extracted feature for the denture to the trained denture dimension evaluation model to evaluate the production dimensions of the denture, and the output unit operates to output the evaluated production dimensions.
In this application, the "production level" refers to the instructions given by a dentist regarding the production of dentures or learning dentures. Specifically, it refers to the respective contents (data) of the instructions regarding the gap size between the treatment tooth and adjacent teeth, called "contact," the instructions regarding the gap size between the treatment tooth and opposing teeth, called "bite," and the instructions regarding the gap size between the treatment tooth and prosthesis, called "fit." In addition, since it is necessary to indicate the identification of the treatment tooth, adjacent teeth, and opposing teeth using dentition data, for convenience, the "production level" also includes dentition data.

In addition, the denture dimension evaluation system of the second invention is characterized in that, in the first invention, it has a trained model generation unit that generates the denture dimension evaluation model using the training data.
In the denture dimension evaluation system configured as described above, the trained model generation unit operates to generate a trained denture dimension evaluation model using as training data the features formed by adding the production-level dentition data for the training denture and at least one of the three pieces of data: production-level contact data, bite data, and fit data, and the production dimensions that have been assigned in advance as correct labels corresponding to the features of the training denture.

The third invention, a denture dimension evaluation method, is a denture dimension evaluation method for converting a production level for a denture indicated in a dental laboratory instruction sheet prepared by a dentist into production dimensions, and enabling a denture to be produced based on the production dimensions and converted into the production dimensions in accordance with the characteristics of the instruction from the dentist, the method comprising: a feature extraction step of extracting a feature consisting of dentition data of the production level for the denture indicated in the dental laboratory instruction sheet and at least one of three data items of contact data, bite data and fit data of the production level; a denture dimension evaluation step of evaluating the production dimensions for the denture by applying the feature to a denture dimension evaluation model; and an output step of outputting the production dimensions for the denture evaluated in the denture dimension evaluation step, the denture dimension evaluation model being a denture dimension evaluation model trained on a combination of the feature of the production level for a learning denture prepared by the dentist and the production dimensions of the learning denture that have been previously assigned corresponding to the feature of the learning denture as training data.
In the denture dimension evaluation method configured as above, the feature extraction step operates to extract a feature consisting of the dentition data of the production level for the denture entered by the dentist in the dental technician instruction sheet and at least one of the three data of the contact data, bite data, and fit data of the production level. Also, the denture dimension evaluation step operates to input the extracted feature for the denture to the trained denture dimension evaluation model to evaluate the production dimensions of the denture, and the output step operates to output the evaluated production dimensions.

The denture dimension evaluation method of the fourth invention is characterized in that, in the third invention, it has a trained model generation step of generating the denture dimension evaluation model trained using the training data.
In the denture dimension evaluation method configured as above, the trained model generation process operates to generate a trained denture dimension evaluation model using as training data the features formed by adding the production-level dentition data for the training denture and at least one of the three pieces of data, namely production-level contact data, bite data, and fit data, and the production dimensions that have been assigned in advance as correct labels corresponding to the features of the training denture.

The denture dimension evaluation program of the fifth invention is a denture dimension evaluation program executed by a computer to convert the production level for a denture indicated in a dental laboratory instruction sheet prepared by a dentist into production dimensions, and to enable conversion to the production dimensions in accordance with the characteristics of the instruction of the dentist and production of a denture based thereon, and is characterized in that it has a denture dimension evaluation model that has been trained using a combination of feature values formed by adding dentition data of the production level for a learning denture prepared by the dentist and at least one of three data items of contact data, bite data and fit data of the production level, and the production dimensions of the learning denture that have been assigned in advance corresponding to the feature values of the learning denture, as training data, and has a feature value extraction process that extracts the feature values of the production level for the denture indicated in the dental laboratory instruction sheet, a denture dimension evaluation process that evaluates the production dimensions for the denture by applying the feature values to the denture dimension evaluation model, and an output process that outputs the production dimensions for the denture evaluated in the denture dimension evaluation process.
In the denture dimension evaluation program having the above configuration, the feature extraction step operates to extract a feature consisting of the dentition data of the production level for the denture entered by the dentist in the dental technician instruction sheet and at least one of the three data of the contact data, bite data, and fit data of the production level. The denture dimension evaluation step operates to input the extracted feature for the denture to the trained denture dimension evaluation model to evaluate the production dimensions of the denture, and the output step operates to output the evaluated production dimensions.

The denture dimension evaluation program of the sixth invention is characterized in that, in the fifth invention, it has a trained model generation process for generating the denture dimension evaluation model trained using the training data.
In the denture dimension evaluation program configured as above, the trained model generation process operates to generate a trained denture dimension evaluation model using as training data the features formed by adding the production-level dentition data for the training denture and at least one of the three pieces of data, namely production-level contact data, bite data, and fit data, and the production dimensions that have been assigned in advance as correct labels corresponding to the features of the training denture.

In the denture dimension evaluation system, which is the first invention, the feature extraction unit extracts features consisting of the dentition data of the denture production level entered by the dentist in the dental laboratory instruction sheet and at least one of the three subjective data expressed by the dentist himself: contact data, bite data, and fit data, and the evaluation unit inputs the extracted denture feature to a trained denture dimension evaluation model, and can evaluate the objective denture production dimensions according to the dentist's instruction characteristics. Therefore, it is possible to convert the subjective and ambiguous denture feature based on the instruction characteristics for each dentist into objective and clear production dimensions that can be produced in a dental laboratory, making it possible to improve the quality and yield of denture production. Furthermore, it is also possible to promote improvements in the yield and working environment of dental laboratories by reducing requests for corrections and re-production of dentures.

The denture dimension evaluation system, which is the second invention, can generate a trained denture dimension evaluation model, and can further train the generated denture dimension evaluation model using the feature values of the dental laboratory instructions used for evaluation and the actual manufacturing dimensions as training data, making it possible to construct a more accurate denture dimension evaluation model. By constructing a more accurate denture dimension evaluation model, it is possible to further enhance the effects described in the first invention.

The third invention, the denture dimension evaluation method, is an invention that regards the first invention, the denture dimension evaluation system, as a method invention, and therefore the effects of implementing this method invention are the same as those of the first invention.

The fourth invention, the denture dimension evaluation method, is an invention that regards the second invention, the denture dimension evaluation system, as a method invention, and therefore the effects of implementing this method invention are the same as those of the second invention.

The denture dimension evaluation program, which is the fifth invention, is an invention that is considered as a program invention that executes the denture dimension evaluation method, which is the third invention, using a computer, and therefore the effects of implementing this program invention are similar to the effects of the first and third inventions.

The denture dimension evaluation program, which is the sixth invention, is an invention that is considered as a program invention that executes the denture dimension evaluation method, which is the fourth invention, using a computer, and therefore the effects of implementing this program invention are similar to the effects of the second and fourth inventions.

1 is a block diagram of a denture dimension evaluation system according to an embodiment of the present invention. (a) is a conceptual diagram showing human dentition, (b) is a conceptual diagram showing contact, which is one of the features, (c) is a conceptual diagram showing bite, which is one of the features, and (d) is a conceptual diagram showing fit, which is one of the features. FIG. 11 is a flow chart showing the flow for generating a denture dimension evaluation model by learning using a denture dimension evaluation system according to an embodiment of the present invention, and the flow for denture dimension evaluation using the learned denture dimension evaluation model. FIG. 1 is a conceptual diagram of a dental technique instruction sheet prepared by a dentist. FIG. 2 is a conceptual diagram showing the configuration of a denture dimension evaluation model in accordance with the present embodiment in the case of pattern 1. FIG. 13 is a conceptual diagram showing the dentition of the input layer for pattern 1 of the denture dimension evaluation model according to the present embodiment, expressed in numbers based on international standards. FIG. 13 is a conceptual diagram showing the configuration of the denture dimension evaluation model in accordance with the present embodiment in the case of pattern 2. 1 is a conceptual diagram showing a denture dimension evaluation model used in an evaluation unit of a denture dimension evaluation system according to an embodiment of the present invention. FIG.

Fig. 1 is a block diagram of a denture size evaluation system according to an embodiment of the present invention. In Fig. 1, the denture size evaluation system 1 is composed of a processing unit 2 consisting of an input unit 3, a feature extraction unit 4a, a trained model generation unit 4b, an evaluation unit 5 and an output unit 6, a teacher data database 7 connected to the processing unit 2, an input data database 13, an output data database 19 and a trained model database 24.
This denture size evaluation system 1 can be assumed to be a system equipped with an arithmetic circuit and a storage device capable of performing the respective functions within a computer server.
The input unit 3 functions as a data input unit or data receiving unit, and corresponds to a receiving device for data related to educational dental technique instructions and dental technique instructions, etc., or a keyboard, mouse, pen tablet, optical reading device, etc. used by administrators and users of the denture dimension evaluation system 1 when inputting data.
The output unit 6 may be a display device, a transmitting device to an information and communication network, or a transmitter that transmits data to other devices as a data transmitting unit. That is, even if the components in Fig. 1 are not integrated, the input unit 3 and the output unit 6 may be provided separately, and a system configuration may be formed in which data and information are transmitted and received by wire or wirelessly between the processing unit 2, databases 7, 13, 19, 24, or a mobile terminal (not shown) or a viewing terminal (not shown) that is provided as necessary.

In the process of generating denture dimension evaluation models 25a, 25b as trained models, the feature extraction unit 4a of the processing unit 2 functions as a device that extracts features consisting of dentition data and at least one of the three data items of contact data, bite data, and fit data from the production level content described in the training dental technician instructions input from the input unit 3.

Here, the dentition data, contact data, bite data and fit data will be described with reference to Fig. 2(a) to (d). Fig. 2(a) is a conceptual diagram showing a human dentition, (b) is a conceptual diagram showing contact, which is one of the feature quantities, (c) is a conceptual diagram showing bite, which is one of the feature quantities, and (d) is a conceptual diagram showing fit, which is one of the feature quantities.
The dentition indicates the position of the row of human teeth as shown in Fig. 2(a). Humans have eight teeth on each side of the upper dentition 26 and the lower dentition 27. The dentist specifies the location of the teeth by assigning numbers "1" to "8" for the top, bottom, left and right, and then indicates the treatment tooth 30 (Figs. 2(b) to (d)), adjacent tooth 31 (Fig. 2(b)), and opposing tooth 33 (Fig. 2(c)) on the dental technique instruction sheet. Therefore, the dentition data refers to the numbers for specifying the location of the teeth specified by the dentist, and the data indicated by the tooth icon positions.
Next, the contact data refers to instructions regarding the gap size (contact 32) between the treatment tooth 30 and the adjacent tooth 31 as shown in Fig. 2(b), the bite data refers to instructions regarding the gap size (bite 34) between the treatment tooth 30 and the opposing tooth 33 as shown in Fig. 2(c), and the fit data refers to instructions regarding the gap size (fit 36) between the treatment tooth 30 and a cap (prosthesis) 35 as shown in Fig. 2(d). Note that in the case of dentures, not only the treatment tooth 30, which is a filed tooth, but also a mucous membrane is substituted for the treatment tooth 30 in Fig. 2(d) and a denture is substituted for the cap 35, and the fit 36 expresses the wearing comfort.

Returning to Figure 1, in the operation process of the denture dimension evaluation models 25a, 25b, they function as a device that extracts features consisting of dentition data and at least one of the three data items of contact data, bite data, and fit data from the production level contents described in the dental technician instructions input from the input unit 3 for the production of dentures to be used in treatment.
The trained model generating unit 4b is a device that generates a trained denture dimension evaluation model using, as training data, feature quantities consisting of the dentition data at the production level for the training denture extracted from the training dental technician instruction sheet, and at least one of the three data items of contact data, bite data, and fit data at the production level, and the production dimensions previously assigned as correct labels corresponding to the feature quantities of the training denture. Hereinafter, the pair of data consisting of the contact data, bite data, and fit data of the feature quantities of the training denture in the training data and the production dimensions previously assigned as correct labels corresponding to each of them will be referred to as contact training data, bite training data, and fit training data, respectively.
The evaluation unit 5 is a device that evaluates the manufacturing dimensions of a denture by applying the feature amounts extracted from the dental technician instructions for denture manufacturing extracted by the feature amount extraction unit 4a to a denture dimension evaluation model.

The teacher data database 7 is a database which readably stores the following: dentition data 9a-9c at the production level for learning dentures extracted from a learning dental technician instruction sheet; contact teacher data 10a-10c consisting of contact data at the production level and contact data of production dimensions which is assigned in advance as a corresponding correct label; byte teacher data 11a-11c consisting of byte data at the production level and byte data of production dimensions which is assigned in advance as a corresponding correct label; and fit teacher data 12a-12c consisting of fit data at the production level and fit data of production dimensions which is assigned in advance as a corresponding correct label.
The reason why the symbols "a" to "c" are added is because it is assumed that there are three dentists, "a" to "c", when the denture size evaluation system 1 according to this embodiment is operated, and teacher data sets 8a to 8c are stored for each of the dentists. As described above, since the denture size evaluation system is intended to reduce variation due to the individuality of each dentist, it is desirable to store a data set for each dentist in the teacher data database 7. Therefore, even if there are dentists other than the three, it is desirable to operate the system by storing the data as a data set for each dentist.

Next, the input data database 13 is a database that readably stores the dentition data 15a-15c at the production level for dentures, the contact data 16a-16c at the production level, the bite data 17a-17c at the production level, and the fit data 18a-18c at the production level, which are extracted from the dental technician instructions when using the denture dimension evaluation system 1. In this input data database 13 as well, three input data sets 14a-14c are stored, assuming three dentists.

Finally, the output data database 19 is a database in which the evaluation unit 5 reads out the denture dimension evaluation models 25a, 25b generated by the trained model generation unit 4b, inputs the input data sets 14a-14c stored in the input data database 13 to this, and stores the data resulting from the evaluation of the manufacturing dimensions of the denture to be created based on the dental laboratory instructions in a readable manner. Specifically, similar to the teacher data database 7 and the input data database 13, three dentists are assumed, and in addition to the contact evaluation data 21a-21c, bite evaluation data 22a-22c, and fit evaluation data 23a-23c, the same dentition data 15a-15c as at the time of input are also stored in a readable manner as the output data sets 20a-20c.

Next, with reference to Figures 3 to 5A and 5B in addition to Figure 1, the flow of generating denture dimension evaluation models 25a, 25b in the denture dimension evaluation system 1 and denture dimension evaluation using the trained denture dimension evaluation models 25a, 25b will be explained.
Fig. 3 is a flow diagram showing the flow of generating a denture dimension evaluation model by learning executed by a denture dimension evaluation system according to an embodiment of the present invention, and the flow of denture dimension evaluation using the learned denture dimension evaluation model. Fig. 3 also shows the execution process for the denture dimension evaluation method and denture dimension evaluation program of the present invention, and explaining the flow of generating a denture dimension evaluation model in the denture dimension evaluation system 1 and denture dimension evaluation using the learned denture dimension evaluation model while referring to this figure is equivalent to explaining the embodiment of the denture dimension evaluation method and denture dimension evaluation program using a computer.
In FIG. 3, the components connected to the symbols indicated by leading lines from the process descriptions indicated by "S" and the dashed lines covering the process descriptions are components of the denture dimension evaluation system 1 shown in FIG. 1, and are given the same symbols.

3, step S1 of the process of generating a trained model on the upper side is a process of inputting information on the feature amounts indicated in the dental laboratory instructions for the training denture, i.e., dentition data, contact data at the production level, bite data at the production level, and fit data at the production level, from the input unit 3. The information on the feature amounts in this step S1 is a concept that includes the above data itself or dental laboratory instructions that include those data.
Furthermore, when inputting information related to the feature quantities into the input unit 3, in the case of inputting the data itself, it is preferable to input based on the letters and numbers representing the dentition data and the numbers (indexes) such as 1 to 7 representing the production levels of the contact data, bite data, and fit data, which are written in the dental technique instruction sheet, which will be described later with reference to Figure 4. On the other hand, when the input unit 3 is a scanner device, information other than the feature quantities may also be input by scanning the dental technique instruction sheet itself.

Here, the dental laboratory instruction sheet will be described with reference to Fig. 4. Fig. 4 is a conceptual diagram of a dental laboratory instruction sheet prepared by a dentist.
4, the dental technique instruction sheet 40 has a name/clinic name display field 41 and a dental clinic address display field 42, and shows information about the doctor who prepared the dental technique instruction sheet 40. Icons related to the upper dentition 26 and the lower dentition 27 are shown together with numbers indicating the positions of the teeth. For example, if the hatched tooth indicated with the number "5" on the "lower left" (lower right in the figure) is the treatment tooth 30, the dentist will show in the dentist check 48 that the tooth is the "5th" tooth on the "lower left" dentition. In this embodiment, the dentist check 48 is shown on the "lower left" dentition and the number "5" indicating the tooth number 43 as an entry by the dentist, but the dentist check 48 may also specify the icon of the tooth shape below that or the tooth number 43 added to the tooth shape. In this embodiment, the dentist check 48 is represented as a graphic such as the katakana character "re", however, the actual shape of the dentist check 48 is not limited to a graphic such as "re", and may be any graphic, number, letter or symbol that identifies a tooth.
Shown on the left side of the area showing the dentition are a contact entry field 45, a bite entry field 46, and a fit entry field 47. These are written as indicators 44 with letters showing the degree of "loose,""normal," or "tight," and numbers "1" to "7" corresponding to the degree and a "scale," and the dentist expresses the contact data, bite data, or fit data of the production level by writing a dentist's check 48 on the "scale." When writing the indicators 44, the dentist's subjective indication characteristics are revealed.
The dental technique instruction sheet 40 is common to both learning dentures and treatment dentures. The dental technique instruction sheet 40 is not limited to the one shown in Fig. 4, but may display at least one of the following: the name of the dentist, instructions regarding the dentition, and instructions regarding the contact data, bite data, and fit data for production. For example, while showing the indicators 44, the dentist's entry may be shown in numbers instead of the dentist's check 48.

Returning to Fig. 3, step S2 is a process in which the feature extraction unit 4a extracts the feature of the production level from the information on the production level input in step S1 via the input unit 3. If the information in step S1 is the data on the production level itself (hereinafter, this case will be referred to as pattern 1), the feature of the production level is extracted from the information as a combination of dentition data, contact data, bite data, and fit data of the required production level.
Furthermore, if the information input in step S1 is the entire dental technique instruction 40 or a part of it (hereinafter, this case will be referred to as pattern 2), the feature amounts as a combination of the dentition data, contact data, bite data, and fit data of the required production level are read and extracted from it. In this case, the extraction of the production level feature amounts can be performed by using the feature amount extraction unit 4a as a device equipped with an OCR (optical character recognition) function and an image recognition function, which can read the letters, numbers, figures, or symbols written on the dental technique instruction 40.
The extracted features of the production level of the training denture are readably stored in the teacher data database 7 by the feature extraction unit 4a as dentition data 9a to 9c, contact teacher data 10a to 10c, bite teacher data 11a to 11c and fit teacher data 12a to 12c.

In step S3, the manufacturing dimensions of the learning denture as the correct label are input via the input unit 3, which readably stores them as contact learning data 10a-10c, bite learning data 11a-11c, and fit learning data 12a-12c in the learning data database 7. As described above, each learning data represents a pair of data consisting of the feature amount of the learning denture's manufacturing level and the manufacturing dimensions as the correct label, and is therefore stored as a pair of data.
Step S4 is a process in which the trained model generating unit 4b reads out the contact teacher data 10a to 10c, the part-time teacher data 11a to 11c, and the fit teacher data 12a to 12c from the teacher data database 7 and performs machine learning.

Here, the machine learning process by the trained model generating unit 4b in step S4 will be described with reference to Figures 5A and 5B. Figure 5A is a conceptual diagram showing the configuration of the denture size evaluation model in accordance with the present embodiment in the case of pattern 1, and Figure 5B is a conceptual diagram showing the dentition of the input layer in the case of pattern 1 of the denture size evaluation model when expressed in numbers based on international standards.
In Figure 5A, the input layer 50a of the denture dimension evaluation model 25a contains, in addition to the dentition data 9a extracted from the dental technician instructions 40 for the learning denture, contact data 53a, bite data 54a and fit data 55a at the production level described by the dentist in the dental technician instructions 40, from among the contact teacher data 10a, bite teacher data 11a and fit teacher data 12a.
In pattern 1, the data is extracted, and the contact data 53a is 2.3, the bite data 54a is 4.5, and the fit data 55a is 4.8. These numbers are directly written by the dentist in the dental technique instruction sheet 40, or are previously extracted in some way from the dentist's check 48 attached to the displayed scale or the like and expressed as numerical values, and have no unit.

On the other hand, in the output layer 52a, the dentition data 9a is output as it is even after passing through the intermediate layer 51a from the input layer 50a, but for the contact data 53a of 2.3, the byte data 54a of 4.5, and the fit data 55a of 4.8 of the production level for the learning denture input to the input layer 50a, the contact production dimension 53b of 0.23 mm, the byte production dimension 54b of 0.05 mm, and the fit production dimension 55b of 0.012 mm of the correct label are output as the production dimensions of the learning denture. These correct label production dimensions are objective and quantitative production dimensions obtained in response to the contact data 53a, etc. of the production level obtained according to the dentist's instruction characteristics. These constitute a pair of contact teacher data 10a, byte teacher data 11a, and fit teacher data 12a.
In addition, in FIG. 5A, the dentition data 9a is shown as data with a dentist's check 48 attached to the tooth number 43 and the position of the dentition. However, these may be assigned numbers based on international standards, with "upper right" being 1, "upper left" being 2, "lower left" being 3 and "lower right" being 4, and together with the tooth numbers 43 which are 1 to 8 from the front teeth, for example, the fourth tooth from the lower right front tooth indicated by the dentist's check 48 in the dentition of FIG. 5A may be expressed as "44" as shown in FIG. 5B. This makes it possible to treat the input in the input unit 3 and the extraction in the feature extraction unit 4a as numerical data itself.
In this case, it is advisable to instruct the dentist to write numbers based on the international standard together with the tooth number 43 in the dentition display field of the dental technician instruction sheet 40 shown in Fig. 4. In this case, the number may be written together with the tooth number 43, such as "upper right (1)", and the dentist may also check 48. In addition, when "44" is input in the dentition data 9a of the input layer 50a, the dentition data 9a is also output as "44" in the output layer 52a of Fig. 5A.
In this way, the dentition data 9a is output as is from the input layer 50a and is not input to the denture dimension evaluation model function 56a in the intermediate layer 51a. Therefore, as long as the position of the treatment tooth 30 in the dentition is identified and there are no problems with the input unit 3 or feature extraction unit 4a, there is no need to be concerned with the format of the data.

In the intermediate layer 51a, a denture dimension evaluation model function 56a is arranged to perform machine learning so that the contact manufacturing dimensions 53b of the learning denture as the correct label output in the output layer 52a are derived from the contact data 53a of the manufacturing level of the learning denture input in the input layer 50a. When the contact data 53a of the manufacturing level is expressed numerically in the input layer 50a as in pattern 1, the model function is preferably a function obtained as a result of a regression analysis using the least squares method or the like, but the type of the denture dimension evaluation model function 56a is not particularly limited, and any function other than a linear function may be used as long as the causal relationship or correlation between the contact data 53a as the input value and the contact manufacturing dimensions 53b as the output value is clarified.
Therefore, in the case of pattern 1, the machine learning process by the trained model generation unit 4b in step S4 is a process of performing regression analysis to find a function that obtains the causal relationship or correlation between the production level contact data 53a, etc. as input values and the contact production dimensions 53b of the correct label, etc. as output values.
In this way, the trained model generation unit 4b generates a trained denture dimension evaluation model function 56a through regression analysis, etc., for receiving input of contact data 53a at the production level of the training denture and outputting contact production dimensions, etc., and uses this as the intermediate layer 51a.The trained denture dimension evaluation model 25a, including the preceding and following input layers 50a and output layer 52a, is stored readably in the trained model database 24.

Next, with reference to Fig. 6, the machine learning process by the trained model generating unit 4b in step S4 will be described for the case of pattern 2. Fig. 6 is a conceptual diagram showing the configuration of the denture dimension evaluation model according to the present embodiment for the case of pattern 2.
In the denture dimension evaluation model 25b of Figure 6, what differs in the case of pattern 2 from pattern 1 is that what is input by the input unit 3 and what is extracted by the feature extraction unit 4a is not numerical data, but the learning dental technician instruction sheet 40 itself, which is an image of the indicator 44 and dentist check 48 in the dental technician instruction sheet 40 as shown in the input layer 50b, and this is input as the contact data image 57a, bite data image 58a and fit data image 59a of the production level of the learning denture.
In the case of pattern 2, as shown in the input layer 50b of Fig. 6, even if the dentition data 9a includes a dentist check 48 for the dentition position and tooth number 43, there is no problem because the input unit 3 and feature extraction unit 4a are equipped with scanning and character and graphic recognition functions. Also, as in pattern 1, there is no need to input the denture size evaluation model function 56b in the intermediate layer 51b.
In the output layer 52b, the tooth row data 9a is output as is, as in pattern 1, and the contact manufacturing dimensions 53b and the like as the correct answer label are also the same as in pattern 1.
A possible machine learning technique for pattern 2 is a denture dimension evaluation model function 56b using a neural network that inputs the index 44 in the dental technician instruction sheet 40 and the contact data image 57a, bite data image 58a, and fit data image 59a of the dentist check 48 as shown in the input layer 50b of Figure 6, and outputs the denture contact manufacturing dimensions 53b, bite manufacturing dimensions 54b, and fit manufacturing dimensions 55b.

In this case, the machine learning process inputs the indicators 44 in the dental technician instructions 40 and the contact data image 57a of the dentist check 48, compares the output denture contact manufacturing dimensions with the correct label contact manufacturing dimensions 53b, and optimizes the parameters of the denture dimension evaluation model function 56b using the neural network of the intermediate layer 51b so that the output denture contact manufacturing dimensions approach the correct label contact manufacturing dimensions 53b. Examples of these parameters include weights (coupling coefficients) between neurons and coefficients used in the activation function. The optimization method for these parameters is performed using, for example, the backpropagation method.
In this way, the trained model generation unit 4b generates a trained denture dimension evaluation model function 56b using a neural network that receives inputs such as the indicators 44 in the dental technician instructions 40 and the contact data image 57a of the dentist's check 48 and outputs contact manufacturing dimensions, etc., and uses this as the intermediate layer 51b.The trained denture dimension evaluation model 25b, including the preceding and following input layers 50b and output layer 52b, is stored readably in the trained model database 24.
In this embodiment, a convolutional neural network (CNN) that is suitable for image processing is used as the neural network. However, other neural networks constructed using other learning algorithms, such as a support vector machine (SVM), may also be used as long as the correct correlation between the input and output can be maintained.

In addition, in Figures 5A, 5B and 6, the production level data by dentist "a" is used as the input, and further, the three data of dentition data, production level contact data, bite data and fit data are all used as feature quantities, but the data related to the production level may include at least one of the three data. This is because in the trained model generation unit 4b in step S4, the production level contact data, bite data and fit data in the feature quantities are trained independently for each of the dentition data. Furthermore, the number of target dentists may be increased beyond "a" to include "b" or "c" or others, but in order to reflect the individuality of each dentist, the trained model generation unit 4b needs to train independently for each dentist.

3, step S5 is a process in which the trained model generation unit 4b obtains trained denture dimension evaluation models 25a, 25b through the machine learning executed in step S4. The obtained denture dimension evaluation models 25a, 25b are stored in the trained model database 24 so as to be readable by the trained model generation unit 4b.
Up to this step S5, trained denture dimension evaluation models 25a, 25b are obtained, and the trained model generation process is completed.

Next, the operation process of the trained model will be described with reference to Fig. 7 in addition to Fig. 3. Fig. 7 is a conceptual diagram showing a denture size evaluation model operated by the evaluation unit of the denture size evaluation system according to an embodiment of the present invention. In Fig. 7, both the case of pattern 1 and the case of pattern 2 are described as the dentition data 15a-15c, the contact data 16a-16c, the bite data 17a-17c, and the fit data 18a-18c shown in the input layer 50c of the denture size evaluation models 25a, 25b.
In the input layer 50c, in pattern 1, the dentition data 15a-15c indicating that the lower right tooth number 43 is tooth "4", the contact data 16a-16c of "3.2", the bite data 17a-17c of "3.7", and the fit data 18a-18c of "4.6" are input as extracted values, while in pattern 2, an image is input in which the data is extracted from the dental technique instruction sheet 40 and a dentist check 48 having an indicator 44 and reflecting the dentist's instruction characteristics is indicated. Note that the denture size evaluation model 25a is for pattern 1, and the denture size evaluation model 25b is for pattern 2.
In Fig. 7, the dentition data 15a to 15c are shown as data with the tooth number 43 and the position of the dentition marked with a dentist check 48, but they may be expressed as numerical data, as in the case of the dentition data 9a described with reference to Figs. 5A and 5B. In this way, the input in the input unit 3 and the extraction in the feature extraction unit 4a can be handled as numerical data itself. However, since the dentition data 15a to 15c are output as they are from the input layer 50c and are not input to the denture size evaluation model functions 56a and 56b in the intermediate layer 51c, the format of the data is not important as long as the position of the treatment tooth 30 in the dentition is specified and there is no problem in the input unit 3 or the feature extraction unit 4a.
3, step S6 is a step of inputting the dental technique instruction sheet 40 for the denture. The data on the production level may be extracted from the dental technique instruction sheet 40 as numerical data in some way as in pattern 1, or the description of the dental technique instruction sheet 40 itself may be input as in pattern 2. The input is made to the input unit 3.

Step S7 is a process in which the feature extraction unit 4a extracts the feature of the production level from the information on the production level input via the input unit 3 in step S6. If the information in step S6 is pattern 1, the feature of the production level is extracted as a combination of at least one of the three data of the dentition data 15a-15c, the contact data 16a-16c, the bite data 17a-17c, and the fit data 18a-18c of the production level of the required denture from the information. Note that if only the numerical data required for the production dimensions is input in advance in step S6, the input numerical data is extracted as it is in step S7.
On the other hand, if the information entered in step S6 is pattern 2, then the dentition data 15a to 15c at the required denture production level from the dental technician instructions 40, the contact data image 57b as contact data 16a to 16c, the bite data image 58b as bite data 17a to 17c, and the fit data image 59b as fit data 18a to 18c are read and extracted as an image relating to the features as a combination of at least one of these three data and the dentition data 15a to 15c.
The feature extraction unit 4a readably stores the extracted features in the input data database 13 as input data sets 14a to 14c, that is, dentition data 15a to 15c, contact data 16a to 16c, bite data 17a to 17c, and fit data 18a to 18c, in both cases of pattern 1 and pattern 2.

Step S8 is a step of inputting the feature quantities extracted in step S7 to the denture dimension evaluation models 25a, 25b. In this step, the evaluation unit 5 reads out one of the denture dimension evaluation models 25a, 25b from the trained model database 24 according to the pattern of the feature quantities extracted by the feature quantity extraction unit 4a. Furthermore, the evaluation unit 5 reads out and inputs a data set as shown as pattern 1 in FIG. 7 from the input data database 13 to the input layer 50c of the denture dimension evaluation models 25a, 25b in the case of pattern 1, and reads out and inputs a data set as shown as pattern 2 from the input data database 13 in the case of pattern 2.
In addition, in FIG. 7, data sets for both patterns 1 and 2 are shown in the input layer 50c, and a data set including all of the contact data 16a to 16c, bite data 17a to 17c, and fit data 18a to 18c are shown, but the extracted feature may be a combination of at least one of the three data related to the production level, such as the contact data 16a to 16c, for the dentition data 15a to 15c required as the production dimensions, and although it appears that the data of three dentists, designated by the symbols "a" to "c", are input simultaneously, the data of one dentist may be input one at a time, or multiple dentists may be input simultaneously.

Step S9 is a process for evaluating the denture dimensions, i.e., the manufacturing dimensions of the denture. In this process, the evaluation unit 5 evaluates the manufacturing dimensions of the denture using the denture dimension evaluation model functions 56a and 56b in the intermediate layer 51c according to the feature values input to the input layer 50c. In other words, this is a process for evaluating the manufacturing dimensions of the denture objectively and quantitatively according to the feature values reflecting the subjective and qualitative instruction characteristics of the dentist.
Although it has been explained that data for multiple dentists may be input simultaneously in step S8, since the denture dimension evaluation model functions 56a, 56b are trained using teacher data for each dentist, it is necessary to use an individual function for each individual dentist.
Step S10 is a step of outputting the manufacturing dimensions of the dentures. In this step, the evaluation unit 5 outputs a set of the dentition data 15a-15c, the contact evaluation data 21a-21c, the bite evaluation data 22a-22c, and the fit evaluation data 23a-23c to the output layer 52c as the manufacturing dimensions of the dentures evaluated by the denture dimension evaluation model functions 56a, 56b in the intermediate layer 51c, and further the evaluation unit 5 stores these data in the output data database 19 as the output data sets 20a-20c in a readable manner.
At the dental laboratory, the necessary output data sets 20a to 20c can be read out from the output data database 19 via the output section 6 of the processing section 2, and dentures can be produced based on the objective, specific and quantitatively obtained denture production dimensions.

Therefore, in the denture size evaluation system 1 according to the present embodiment, it is possible to convert the subjective, qualitative, and ambiguous characteristic quantities of the dentures, which are reflected in the dental laboratory instructions for each dentist, into objective, quantitative, and clear manufacturing dimensions that can be manufactured at the dental laboratory, and it is possible to improve the quality and increase the yield of dentures manufactured. Furthermore, it is possible to contribute to improving the work efficiency of dental laboratories by reducing requests from dentists for corrections and remanufacturing of dentures, and it is also possible to promote the improvement of the working environment.
Furthermore, the denture size evaluation system 1 includes the trained model generation unit 4b and the teacher data database 7, so that the teacher data sets 8a-8c can be updated over time when generating the denture size evaluation models 25a, 25b. Therefore, it is possible to build a trained model with higher accuracy by updating the denture size evaluation models 25a, 25b using the trained model generation unit 4b in accordance with the accumulation of the dentition data 9a-9c, the contact teacher data 10a-10c, the bite teacher data 11a-11c, and the fit teacher data 12a-12c, and the trained model can be further improved.
In addition, a denture dimension evaluation method in which the denture dimension evaluation system 1 is regarded as a method invention, and a denture dimension evaluation program in which the denture dimension evaluation system 1 is regarded as a program invention, can also achieve effects similar to those that can be achieved by the denture dimension evaluation system 1.

The present invention can be used in a denture dimension evaluation system, method, and program that can evaluate the denture dimensions required when making a denture at a denture lab based on dental laboratory instructions exchanged between a dentist and a dental laboratory.

1... Denture dimension evaluation system 2... Processing unit 3... Input unit 4a... Feature extraction unit 4b... Trained model generation unit 5... Evaluation unit 6... Output unit 7... Teacher data database 8a to 8c... Teacher data set 9a to 9c... Dentition data 10a to 10c... Contact teacher data 11a to 11c... Byte teacher data 12a to 12c... Fit teacher data 13... Input data database 14a to 14c... Input data set 15a to 15c... Dentition data 16a to 16c... Contact data 17a to 17c... Byte data 18a to 18c... Fit data 19... Output data database 20a to 20c... Output data set 21a to 21c... Contact evaluation data 22a to 22c... Byte evaluation data 23a to 23c... Fit evaluation data 24... Trained model database 25a, 25b... Denture dimension evaluation model 26...Upper dentition 27...Lower dentition 30...Treated tooth 31...Adjacent tooth 32...Contact 33...Opposing tooth 34...Bite 35...Cover 36...Fit 40...Dental technique instruction 41...Name/clinic name display field 42...Location display field 43...Tooth number 44...Index 45...Contact entry field 46...Bite entry field 47...Fit entry field 48...Dentist check 50a-50c...Input layer 51a-51c...Intermediate layer 52a-52c...Output layer 53a...Contact data 53b...Contact manufacturing dimensions 54a...Bite data 54b...Bite manufacturing dimensions 55a...Fit data 55b...Fit manufacturing dimensions 56a, 56b...Denture dimension evaluation model function 57a, 57b...Contact data image 58a, 58b...Bite data image 59a, 59b...Fit data image

Claims (4)

A denture size evaluation system for converting a manufacturing level for a denture indicated in a dental laboratory instruction sheet prepared by a dentist into manufacturing dimensions, and enabling a reduction to the manufacturing dimensions according to the instruction characteristics of the dentist and a denture manufacturing based thereon,
a denture size evaluation model that is trained using a combination of the following as training data: a feature value that is formed by adding at least one of the three data items of the dentition data of the training denture made by the dentist and the contact data, bite data, and fit data of the training denture, and a manufacturing dimension that is given in advance in correspondence with the feature value of the training denture;
a feature extraction unit that extracts the feature of the production level for the denture indicated in the dental technician instruction sheet;
an evaluation unit that evaluates the manufacturing dimensions of the denture by applying the feature amounts extracted by the feature amount extraction unit to the denture dimension evaluation model;
and an output unit that outputs the manufacturing dimensions for the denture evaluated by this evaluation unit. The denture size evaluation system according to claim 1, further comprising a trained model generation unit that generates the denture size evaluation model using the training data. A denture size evaluation program executed by a computer to convert a manufacturing level for a denture indicated in a dental laboratory instruction prepared by a dentist into manufacturing dimensions, and to enable reduction to the manufacturing dimensions according to the instruction characteristics of the dentist and denture manufacturing based thereon,
a denture size evaluation model that is trained using a combination of the following as training data: a feature value that is a sum of dentition data of the training denture made by the dentist and at least one of three data items of contact data, bite data, and fit data of the training denture made by the dentist at the training denture level; and a training dimension of the training denture that is previously assigned in correspondence with the feature value of the training denture;
a feature extraction step of extracting the feature of the production level for the denture indicated in the dental technician instruction sheet;
a denture dimension evaluation step of evaluating the manufacturing dimensions of the denture by applying the feature amount to the denture dimension evaluation model;
and an output step of outputting the manufacturing dimensions for the denture evaluated in this denture dimension evaluation step. The denture dimension evaluation program according to claim 3, further comprising a trained model generation step for generating the denture dimension evaluation model trained using the training data.

Images