KR102536834B1 - Method for numbering tooth number and generating interpretation of tooth state using artificial intelligence - Google Patents

Abstract

Disclosed is a method for identifying tooth number by using artificial intelligence, which may automatically identify the tooth number from a panorama tooth image. To this end, the method for identifying tooth number by using artificial intelligence may comprise: a step in which an information processing device receives a panorama tooth image; a step in which the information processing device uses a first object detection module which has learned to divide individual teeth to divide individual teeth in the panorama tooth image to calculate location information with respect to each tooth; a step in which the information processing device uses a second object detection module, which has learned to calculate a cervical line, to calculate the cervical line in the panorama tooth image before calculating location information of the cervical line; a step in which the information processing device compares the location information of the tooth to the location information of the cervical line to identify whether the tooth is an upper jaw tooth or a lower jaw tooth; and a step in which the information processing device uses a first image dividing model, which has learned to divide the dental formula of the individual teeth, to calculate the probability in which the each tooth, sorted as the upper jaw tooth or the lower jaw tooth, may be matched with the dental formula before the tooth number is applied.

Description

Method for numbering tooth number and generating interpretation of tooth state using artificial intelligence}

The present invention relates to a method for identifying tooth numbers and a method for generating a tooth condition interpretation statement using artificial intelligence. In particular, tooth numbers are automatically identified without the need to write by hand, and tooth information or implant information is generated for each identified tooth number to generate a tooth condition interpretation statement. It relates to a tooth number identification method using artificial intelligence capable of automatically generating and a method of generating a tooth condition reading statement.

In dentistry, various pictures such as X-Ray images, CT images, oral images, and facial photographs are required to confirm diagnosis, treatment, and treatment progress. In particular, in the case of orthodontic treatment, temporomandibular joint treatment, and implant treatment, panoramic dental images such as X-ray images of the entire tooth, head radiographs, etc. need this

Conventionally, it is inconvenient for a doctor to manually take these various pictures, attach a plurality of pictures to a patient chart, and assign a tooth number to a panoramic tooth image. That is, the doctor classifies the upper and lower teeth by looking at the photographed panoramic picture and assigns a tooth number, which is inconvenient and may incorrectly assign a tooth number when a tooth is missing in the middle. In addition, by attaching such a panoramic picture to a patient's chart in a conventional dentistry or separately writing and managing medical records, treatment records, implant information, etc., not only is there inconvenience in management, but also doctors and patients can intuitively And there is a problem that it is difficult to intuitively grasp the past treatment records.

The problem to be solved by the present invention is to provide a tooth number identification method using artificial intelligence that can automatically identify a tooth number without the need to write by hand.

Another problem to be solved by the present invention is a tooth condition using artificial intelligence that can automatically generate a tooth condition statement by automatically identifying tooth numbers and generating tooth information or implant information for each identified tooth number without the need to write by hand. It is to provide a method for generating read statements.

A tooth number identification method using artificial intelligence for automatically identifying a tooth number from a panoramic tooth image according to an embodiment of the present invention for achieving the above object includes the steps of receiving a panoramic tooth image in an information processing device, the information In a processing device, by using a first object detection module learned to distinguish individual teeth, discriminating individual teeth from the panoramic tooth image and calculating positional information for each tooth, in the information processing device, determining the alveolar line Calculating an alveolar line from the panoramic tooth image using a second object detection module learned to calculate and calculating the position information of the alveolar line, in the information processing device, the position information of the tooth and the position of the alveolar line Comparing information to distinguish whether the tooth is an upper or lower tooth, and in the information processing device, using a first image classification model learned to distinguish the tooth formula of an individual tooth, the upper or lower teeth are classified into A step of assigning a tooth number by calculating a probability that each tooth matches the tooth formula may be included.

The calculating of the location information of the tooth may include: in the information processing device, using the first object detection module to identify individual teeth in the panoramic tooth image; and in the information processing device, each of the classified teeth. and calculating a first vertical maximum value and a first vertical minimum value as coordinates, and calculating the location information of the alveolar line, in the information processing device, using the second object detection module. Calculating the alveolar lines of the teeth in the panoramic tooth image, and in the information processing device, calculating the second vertical maximum value as coordinates using upper alveolar lines among the calculated alveolar lines, and calculating the calculated alveolar lines A step of calculating the second vertical minimum value as coordinates using alveolar lines of the middle lower portion may be included.

The step of distinguishing whether the tooth is an upper or lower tooth may include, in the information processing device, when the first vertical maximum value is greater than or equal to the second vertical maximum value and the first vertical minimum value is greater than or equal to the second vertical minimum value, the tooth determining as an upper jaw tooth and determining, in the information processing device, as a lower jaw tooth when the first vertical minimum value is less than or equal to the second vertical minimum value and the first vertical maximum value is less than or equal to the second vertical maximum value can include

The step of discriminating whether the tooth is an upper or lower tooth may include, in the information processing device, the second vertical maximum value being between the first vertical maximum value and the first vertical minimum value or the first vertical minimum value being the first vertical minimum value. 2 If the vertical maximum value is greater than or equal to, determining the tooth as an upper jaw tooth, and in the information processing device, the second vertical minimum value is between the first vertical maximum value and the first vertical minimum value, or the first vertical maximum value is the first vertical minimum value. The method may include determining that the tooth is a mandibular tooth when it is equal to or less than the second vertical minimum value.

The step of discriminating whether the tooth is an upper or lower tooth may include, in the information processing device, a case where the first vertical maximum value is greater than or equal to the second vertical maximum value and the first vertical minimum value is less than or equal to the second vertical minimum value; When the first vertical maximum value is less than the second vertical maximum value and the first vertical minimum value is greater than the second vertical minimum value, the distance between the first vertical minimum value and the second vertical maximum value is the first vertical minimum value. If it is smaller than the distance between the maximum value and the second vertical minimum value, it is determined as an upper jaw, and the distance between the first vertical minimum value and the second vertical maximum value is the distance between the first vertical maximum value and the second vertical minimum value. If it is larger than the lower jaw, a step of determining the lower jaw may be further included.

The step of assigning the tooth number may include generating, in the information processing device, maxillary tooth groups in all cases to which the tooth number can be applied according to the number of teeth determined as the upper jaw, in the information processing device, the upper jaw Calculating a probability that each tooth matches a tooth formula by using the first image classification model for each tooth group, a score calculated using the calculated probability for each tooth group of the upper teeth in the information processing device Giving tooth numbers of teeth corresponding to the upper teeth to the tooth numbers of the upper jaw tooth group selected using Generating a lower jaw tooth group, calculating a probability that each tooth matches a tooth formula using the first image classification model for each lower jaw tooth group in the information processing device, and in the information processing device and assigning tooth numbers of teeth corresponding to the lower teeth to the tooth numbers of the lower jaw tooth groups selected using the score calculated using the calculated probability for each lower jaw tooth group.

To achieve the above object, a method for generating a tooth condition reading statement using artificial intelligence for automatically generating a tooth condition reading statement by automatically identifying a tooth number and automatically determining a tooth state from a panoramic tooth image according to an embodiment of the present invention for achieving the other object. In the information processing device, receiving a panoramic tooth image, in the information processing device, by using the first object detection module learned to distinguish individual teeth, individual teeth are distinguished from the panoramic tooth image and assigned to each tooth Calculating location information about the alveolar line in the information processing device, calculating an alveolar line from the panoramic tooth image using a second object detection module learned to calculate an alveolar line and calculating location information of the alveolar line; In the information processing device, comparing the location information of the tooth and the location information of the alveolar line to distinguish whether the tooth is a maxillary tooth or a mandibular tooth; 1 Calculating the probability that each tooth divided into the upper jaw or the lower jaw matches the tooth formula using an image classification model and assigning a tooth number; Using the second image classification model learned to generate state and implant information, it is determined whether it is a tooth or an implant for each tooth number, and when it is determined as a tooth, treatment information for the tooth is generated. Generating implant information related to the implant, and generating a tooth condition reading statement by combining information on whether the tooth is a tooth or an implant for each tooth number, the treatment information, and the implant information in the information processing device. .

A tooth number identification method and a tooth condition interpretation method using artificial intelligence according to an embodiment according to the technical idea of the present invention automatically use artificial intelligence without the need for a doctor to manually assign a tooth number after viewing a panoramic tooth image. By distinguishing the upper and lower teeth and automatically assigning tooth numbers, the surgeon's work burden can be minimized and the tooth numbers can be assigned accurately. In addition, the tooth number identification method and tooth condition statement generation method using artificial intelligence according to an embodiment according to the technical idea of the present invention not only automatically assign tooth numbers, but also use the image of each tooth to determine the corresponding tooth. It automatically detects whether or not you have been treated, what kind of treatment you have received, whether you have had implants, and if you have had implants, what manufacturer and model number you have received, and automatically creates a dental condition report that records such information. Not only can the workload be minimized, but both the doctor and the patient can intuitively judge the current status and past history.

In order to more fully understand the drawings cited in the detailed description of the present invention, a brief description of each drawing is provided.
1 is a flowchart of a tooth number identification method using artificial intelligence according to an embodiment according to the technical idea of the present invention.
FIG. 2 is a view for explaining an embodiment of calculating position information of teeth in FIG. 1 .
FIG. 3 is a diagram for explaining an embodiment of calculating location information of an alveolar line in FIG. 1 .
4 and 5 are views for explaining an embodiment in which tooth numbers are assigned in FIG. 1 .
6 is a flowchart of a method for generating a dental condition reading statement using artificial intelligence according to an embodiment according to another technical concept of the present invention.

In order to fully understand the present invention and its operational advantages and objectives achieved by the practice of the present invention, reference should be made to the accompanying drawings illustrating preferred embodiments of the present invention and the description in the drawings.

Hereinafter, the present invention will be described in detail by describing preferred embodiments of the present invention with reference to the accompanying drawings. Like reference numerals in each figure indicate like members.

1 is a flowchart of a method for identifying tooth numbers using artificial intelligence according to an embodiment according to the technical concept of the present invention, and FIG. 2 is a diagram for explaining an embodiment of calculating tooth position information in FIG. 1 . FIG. 3 is a diagram for explaining an embodiment of calculating position information of the alveolar line in FIG. 1 , and FIGS. 4 and 5 are diagrams for explaining an embodiment of assigning tooth numbers in FIG. 1 . All operations below may be performed in an information processing device, and the information processing device may be various devices such as a computer and a server.

1 to 5, the information processing device may receive a panoramic tooth image (S110). The panoramic tooth image refers to an image or image in which all teeth in the oral cavity are displayed on one image. The panoramic tooth image may be a panoramic X-ray image, but the present invention is not limited to this case, and all other types of images may be included in the panoramic tooth image as long as the entire tooth in the oral cavity can be expressed as one image.

The information processing device may classify individual teeth in the panoramic tooth image using the first object detection module and calculate location information for each tooth (S120). The first object detection module means an object detection module based on deep learning or machine learning, and the first object detection module is a module learned to distinguish an individual tooth (one tooth) from the panoramic image. That is, the first object detection module can distinguish each individual tooth from the panoramic tooth image and recognize the image individually one by one. For example, as shown in FIG. 2, each tooth area 210 or 220) can be recognized as a box shape. The location information may include coordinate values of each of the separated teeth. The coordinate value is a horizontal coordinate value corresponding to the coordinate value of the tooth in the horizontal direction (coordinate value in the x-axis direction) and a vertical coordinate value corresponding to the coordinate value of the tooth in the vertical direction (coordinate value in the y-axis direction) can include In one tooth, the coordinate value located at the top of the vertical coordinate value in the vertical direction is the first vertical maximum value, and the coordinate value located at the bottom in the vertical direction among the vertical coordinate values is defined as the first vertical minimum value. . That is, the first vertical maximum value and the first vertical minimum value may be set for each tooth by the first object detection module. However, even if the coordinate values of each tooth are set in this way, it is impossible to distinguish whether each tooth is a maxillary tooth or a mandibular tooth.

The information processing module may calculate a cervical line from the panoramic tooth image using the second object detection module and calculate location information of the cervical line (S130). The alveolar line is a line connecting the crown of the tooth and means a line marking the boundary between the head and root of the tooth. The second object detection module refers to an object detection module based on deep learning or machine learning, and the second object detection module is a module learned to calculate an alveolar line from the panoramic image. That is, the second object detection module may calculate each alveolar line from the panoramic tooth image. Using the values of the calculated alveolar line, as shown in FIG. 3, the alveolar line area 310 (to the left and right of the center) long rectangular box) can be set. For example, when the alveolar line area 310 is formed as shown in FIG. 3, the upper boundary of the alveolar line area 310 may be determined using upper alveolar lines among the calculated alveolar lines, and the alveolar line area ( 310), the lower boundary may be determined using lower alveolar lines among the calculated alveolar lines. That is, the information processing module calculates the second vertical maximum, which is the upper boundary of the alveolar line area 310, as coordinates using the upper alveolar lines among the calculated alveolar lines, and the lower one of the calculated alveolar lines. A second vertical minimum value, which is the lower boundary of the alveolar line region 310, may be calculated using the meridians as coordinates. For example, the second vertical maximum value may be an average value of upper alveolar lines among the calculated alveolar lines, and the second vertical minimum value may be an average value of lower alveolar lines among the calculated alveolar lines. However, it is not necessary to use the average value of the alveolar lines for the second vertical maximum value and the second vertical minimum value, and may be calculated in various other ways.

Next, the information processing module may compare the location information of the tooth and the location information of the alveolar line to determine whether the tooth is a maxillary tooth or a mandibular tooth (S140). For example, the information processing module may determine the tooth as an upper jaw when the first vertical maximum value is equal to or greater than the second vertical maximum value and the first vertical minimum value is equal to or greater than the second vertical minimum value. In addition, the information processing module may determine the tooth as a mandibular tooth when the first vertical minimum value is equal to or less than the second vertical minimum value and the first vertical maximum value is equal to or less than the second vertical maximum value. As another example, the information processing module determines that the tooth is an upper jaw when the second vertical maximum value is between the first vertical maximum value and the first vertical minimum value or when the first vertical minimum value is greater than or equal to the second vertical maximum value. can do. In addition, the information processing module may determine that the tooth is a mandibular tooth when the second vertical minimum value is between the first vertical maximum value and the first vertical minimum value or when the first vertical maximum value is less than or equal to the second vertical minimum value. there is.

However, when the first vertical maximum value is greater than or equal to the second vertical maximum value and the first vertical minimum value is less than or equal to the second vertical minimum value, and when the first vertical maximum value is less than the second vertical maximum value, the first vertical minimum value is less than the second vertical maximum value. When the vertical minimum value exceeds the second vertical minimum value, it is determined whether the tooth is an upper jaw tooth or a lower jaw tooth. In this case, the information processing module determines that it is an upper jaw when the distance between the first vertical minimum value and the second vertical maximum value is smaller than the distance between the first vertical maximum value and the second vertical minimum value, and the first vertical maximum value If the distance between the minimum value and the second vertical maximum value is greater than the distance between the first vertical maximum value and the second vertical minimum value, it may be determined as a lower jaw tooth.

When the determination of whether the upper or lower teeth is completed with respect to all of the detected teeth while repeatedly performing the same steps as above, the information processing module determines whether the upper or lower teeth are classified using the first image classification model. A tooth number may be assigned by calculating a probability that each tooth matches the tooth formula (S150). The tooth formula is expressed as a formula to express the type and number of teeth in an easy-to-understand manner. In the case of humans, there are 8 tooth formulas on the left/right side of the upper and lower teeth, respectively. That is, formulas 1 and 2 are divided into door teeth (incisors), formula 3 is divided into canines (canines), formulas 4 and 5 are divided into premolars (premolars), and formulas 6, 7, and 8 are molars (back molars). . For example, based on the panoramic tooth image, the teeth on the left side of the maxillary teeth are defined from the middle incisor to the left from tooth formula 1 to tooth formula 8, and tooth formula 1 is tooth number 11, tooth formula 2 is tooth number 12, and tooth formula 3 is Tooth number 13, formula 4 tooth number 14, formula 5 becomes tooth number 15, formula 6 becomes tooth number 16, formula 7 becomes tooth number 17, formula 8 becomes tooth number 18. Similarly, based on the panoramic tooth image, the teeth on the right side of the maxillary teeth are defined from the middle incisors to the right from tooth formula 1 to tooth formula 8. Tooth formula 1 is tooth number 21, tooth formula 2 is tooth number 22, and tooth formula 3 is tooth number 23. , formula 4 tooth number 24, formula 5 tooth number 25, formula 6 tooth number 26, formula 7 tooth number 27, and formula 8 tooth number 28. Based on the panoramic tooth image, the teeth on the left side of the mandibular teeth are defined from the middle incisors to the left from tooth formula 1 to tooth formula 8, with formula 1 being tooth number 31, formula 2 being tooth number 32, formula 3 being tooth number 33, Tooth formula 4 becomes tooth number 34, formula 5 becomes tooth number 35, formula 6 becomes tooth number 36, formula 7 becomes tooth number 37, and formula 8 becomes tooth number 38. Likewise, based on the panoramic tooth image, the teeth on the right side of the mandible are defined from the middle incisor to the right from tooth formula 1 to tooth formula 8. Tooth formula 1 is tooth number 41, tooth formula 2 is tooth number 42, and tooth formula 3 is tooth number 43. , tooth formula 4 tooth number 44, tooth formula 5 tooth number 45, tooth formula 6 tooth number 46, tooth formula 7 tooth number 47, and tooth formula 8 tooth number 48.

The first image classification model refers to an image classification module based on deep learning or machine learning, and the first image classification module is a module learned to classify a tooth type of an individual tooth (one tooth). That is, the first image classification module may assign a tooth number after calculating a probability that the tooth is tooth formula 1 to tooth formula 8 based on the individual tooth image.

The classification of the teeth corresponding to the upper jaw teeth and the teeth corresponding to the lower jaw teeth have been previously completed, and the order of the teeth can be confirmed by aligning them using the horizontal coordinate values of the classified teeth. That is, if the teeth are arranged by arranging the horizontal coordinate values of the teeth corresponding to the upper jaw in a direction from left to right, the order of the teeth corresponding to the upper jaw can be known. Similarly, if the teeth are arranged by arranging the horizontal coordinate values of the teeth corresponding to the lower jaw in a direction from left to right, the order of the teeth corresponding to the lower jaw can be known. However, it is not possible to accurately know the tooth numbers of the teeth arranged in this way. For example, if there are 16 teeth judged as upper teeth, there is a high possibility that the tooth numbers will be assigned in the order of horizontal coordinate values. cannot be accurately assigned.

To this end, the information processing device may generate maxillary tooth groups in all cases to which tooth numbers can be applied according to the number of teeth determined as the maxillary teeth. And the information processing module calculates a probability that each tooth matches the tooth formula for each upper jaw tooth group by using the first image classification model, and calculates a probability that each tooth matches the tooth formula for each upper jaw tooth group. The tooth numbers of the teeth corresponding to the upper teeth may be given as the tooth numbers of the upper teeth group selected using the score. Likewise, in the case of the lower jaw, the information processing device may generate lower jaw tooth groups in all cases to which tooth numbers can be applied according to the number of teeth determined as the lower jaw. And the information processing module calculates a probability that each tooth matches the tooth formula for each lower jaw tooth group using the first image classification model, and calculates a probability that each tooth matches the tooth formula for each lower jaw tooth group. The tooth numbers of the teeth corresponding to the lower teeth may be assigned as the tooth numbers of the tooth group of the lower jaw teeth selected using the score.

For example, when the number of teeth determined to be the upper jaw is 15, the number of teeth that can be the upper jaw tooth group is 16 as shown in FIG. 4 . That is, tooth group 1 with tooth numbers 11 to 17 and 21 to 28, tooth group 2 with tooth numbers 11 to 16, 18, and 21 to 28, and teeth with tooth numbers 11 to 15, 17, 18, and 21 to 28. It can be seen that there are a total of 16 cases when the cases of the maxillary tooth group are divided like Group 3. In addition, the probability that each tooth matches the tooth formula is calculated using the first image classification model for each upper jaw tooth group. For example, as shown in FIG. 5 , the probability that 15 teeth (teeth 1 to 15) that are determined to be upper teeth correspond to tooth formulas 1 to 7, respectively, is determined by the first image classification model. For example, the probability of matching tooth formula for each tooth is determined, such as 80% probability of tooth formula 1, 16% probability of tooth formula 2, and 7% probability of tooth formula 3. In addition, the probability that the teeth match the tooth formula is applied to each tooth group in the order of the teeth. For example, in FIG. 5 , in tooth group 1, tooth number 17 corresponds to tooth formula 7 of tooth 1, so a probability corresponding to tooth formula 7 of tooth 1 is applied to tooth number 17, and tooth number 16 corresponds to tooth formula 6 of tooth 2. Since it corresponds to tooth number 16, the probability corresponding to tooth formula 6 of tooth 2 is applied. When this operation is performed from tooth group 1 to tooth group 16, then, using the score calculated using the probability calculated for each maxillary tooth group, the tooth number of the selected maxillary tooth group is assigned to the maxillary tooth. The tooth number of the corresponding teeth is given.

For example, the scores for each tooth group of the teeth corresponding to the upper teeth or the lower teeth may be defined as in Equation 1 below.

는 i번째 치아(ti)가 i번째 치식(si)일 확률N is the number of teeth of the upper jaw tooth group or the lower jaw tooth group,

is the probability that the ith tooth (ti) is the ith tooth formula (si)

However, the score for each tooth group does not have to be determined by Equation 1, and the score for each tooth group may be determined by another method.

The tooth number of the tooth group having the highest score among the scores determined in this way is eventually determined as the tooth number of the teeth. For example, in the examples of FIGS. 4 and 5 , if tooth group 3 has the highest index number, tooth numbers 18, 17, 15 ... , 26, 27, and 28 are determined for the maxillary teeth without tooth number 16. Likewise, for the lower teeth, tooth numbers of the teeth determined as the lower teeth may be assigned through the same method.

6 is a flowchart of a method for generating a dental condition reading statement using artificial intelligence according to an embodiment according to another technical concept of the present invention. All operations below may be performed in an information processing device, and the information processing device may be various devices such as a computer and a server.

Referring to FIGS. 1 to 6 , in order to perform the method of generating a tooth state reading statement using artificial intelligence, the tooth number identification method using artificial intelligence described with reference to FIGS. 1 to 5 must first be executed. That is, since steps S610 to S650 in FIG. 6 are the same as steps S110 to S150 in FIG. 1 , overlapping descriptions below are replaced with descriptions of steps S110 to S150 in FIG. 1 . After steps S610 to S650 are performed and the tooth number is identified, the information processing device distinguishes between teeth and implants and uses the second image classification model learned to generate information on the state of teeth and implants for each tooth number. It can determine whether it is a tooth or an implant. The second image classification model refers to an image classification module capable of discriminating whether the tooth is a natural tooth or an implant using the image of the individual tooth. The information processing device may generate treatment information of the tooth when the tooth is determined to be a natural tooth, and may generate implant information related to the implant when the tooth is determined to be an implant. In addition, a tooth condition interpretation statement may be generated by combining the information on whether the teeth are teeth or implants, the treatment information, and the implant information for each tooth number (S660). Here, the treatment information may include at least one of information on whether the tooth is treated or not and information on whether or not nerve treatment is performed, and the implant information may include at least one of manufacturer information of the implant and model information of the implant. .

As described above, the optimum embodiment has been disclosed in the drawings and specifications. Although specific terms have been used herein, they are only used for the purpose of describing the present invention and are not used to limit the scope of the present invention described in the claims or defining the meaning. Therefore, those of ordinary skill in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

Claims (15)

A tooth number identification method using artificial intelligence for automatically identifying a tooth number from a panoramic tooth image,
In an information processing device, receiving a panoramic tooth image;
In the information processing device, calculating location information for each tooth by identifying individual teeth in the panoramic tooth image using a first object detection module learned to distinguish individual teeth;
calculating, in the information processing device, an alveolar line from the panoramic tooth image using a second object detection module learned to calculate an alveolar line and calculating location information of the alveolar line;
In the information processing device, comparing the location information of the tooth and the location information of the alveolar line to distinguish whether the tooth is a maxillary tooth or a mandibular tooth; and
In the information processing device, using a first image classification model learned to classify the tooth formula of individual teeth, each tooth divided into the upper jaw or the lower jaw calculates a probability that each tooth matches the tooth formula, and assigns a tooth number contains steps,
Calculating the location information for the tooth,
In the information processing device, discriminating individual teeth from the panoramic tooth image using the first object detection module; and
In the information processing device, calculating a first vertical maximum value and a first vertical minimum value for each of the separated teeth as coordinates,
Calculating the location information of the alveolar line,
calculating, in the information processing device, alveolar lines of the teeth in the panoramic tooth image using the second object detection module; and
In the information processing device, the second vertical maximum value is calculated as coordinates using upper alveolar lines among the calculated alveolar lines, and the second vertical minimum value is calculated as coordinates using lower alveolar lines among the calculated alveolar lines. Including the step of calculating as,
The step of distinguishing whether the teeth are upper or lower teeth,
In the information processing device, comparing the first vertical maximum value, the first vertical minimum value, the second vertical maximum value, and the second vertical minimum value to distinguish whether the tooth is an upper jaw or a lower jaw. Tooth number identification method using artificial intelligence. delete The method of claim 1, wherein the step of distinguishing whether the teeth are upper or lower teeth,
determining, in the information processing device, that the tooth is an upper jaw when the first vertical maximum value is greater than or equal to the second vertical maximum value and the first vertical minimum value is greater than or equal to the second vertical minimum value; and
In the information processing device, if the first vertical minimum value is less than or equal to the second vertical minimum value and the first vertical maximum value is less than or equal to the second vertical maximum value, determining the tooth as a mandibular tooth. Tooth number identification method using intelligence. The method of claim 1, wherein the step of distinguishing whether the teeth are upper or lower teeth,
In the information processing device, determining that the tooth is an upper jaw when the second vertical maximum value is between the first vertical maximum value and the first vertical minimum value or when the first vertical minimum value is greater than or equal to the second vertical maximum value ; and
In the information processing device, determining that the tooth is a mandibular tooth when the second vertical minimum value is between the first vertical maximum value and the first vertical minimum value or when the first vertical maximum value is equal to or less than the second vertical minimum value Tooth number identification method using artificial intelligence, characterized in that it comprises. The method of claim 1, wherein the step of distinguishing whether the teeth are upper or lower teeth,
In the information processing device, the case where the first vertical maximum value is equal to or greater than the second vertical maximum value and the first vertical minimum value is equal to or less than the second vertical minimum value and the first vertical maximum value is less than the second vertical maximum value and when the first vertical minimum value exceeds the second vertical minimum value, when the distance between the first vertical minimum value and the second vertical maximum value is smaller than the distance between the first vertical maximum value and the second vertical minimum value The step of determining the upper jaw teeth and determining the lower jaw teeth when the distance between the first vertical minimum value and the second vertical maximum value is greater than the distance between the first vertical maximum value and the second vertical minimum value Tooth number identification method using artificial intelligence. The method of claim 1, wherein the assigning of the tooth number comprises:
generating, in the information processing device, upper jaw tooth groups in all cases to which tooth numbers can be applied according to the number of teeth determined as the upper jaw teeth;
calculating, in the information processing device, a probability that each tooth matches a tooth formula by using the first image classification model for each tooth group of the upper teeth;
assigning, in the information processing device, tooth numbers of teeth corresponding to the maxillary teeth as the tooth numbers of the maxillary tooth groups selected using the score calculated using the calculated probability for each of the maxillary tooth groups;
generating, in the information processing device, tooth groups of lower jaw teeth in all cases to which tooth numbers can be applied according to the number of teeth determined as the lower jaw teeth;
calculating, in the information processing device, a probability that each tooth matches a tooth formula by using the first image classification model for each tooth group of the lower jaw; and
In the information processing device, assigning tooth numbers of teeth corresponding to the lower jaw teeth as the tooth numbers of the lower jaw tooth groups selected using the score calculated using the calculated probability for each lower jaw tooth group. Tooth number identification method using artificial intelligence, characterized in that. According to claim 6,
The scores for each tooth group of the teeth corresponding to the upper or lower teeth are,

(N is the number of teeth of the upper jaw tooth group or the lower jaw tooth group,

is the probability that the ith tooth (ti) is the ith tooth formula (si)
Tooth number identification method using artificial intelligence, characterized in that calculated as. A method for generating a tooth condition statement using artificial intelligence that automatically identifies a tooth number from a panoramic tooth image, automatically determines a tooth state, and automatically generates a tooth state statement,
In an information processing device, receiving a panoramic tooth image;
In the information processing device, calculating location information for each tooth by identifying individual teeth in the panoramic tooth image using a first object detection module learned to distinguish individual teeth;
calculating, in the information processing device, an alveolar line from the panoramic tooth image using a second object detection module learned to calculate an alveolar line and calculating location information of the alveolar line;
In the information processing device, comparing the location information of the tooth and the location information of the alveolar line to distinguish whether the tooth is a maxillary tooth or a mandibular tooth;
In the information processing device, using a first image classification model learned to classify the tooth formula of individual teeth, each tooth divided into the upper jaw or the lower jaw calculates a probability that each tooth matches the tooth formula, and assigns a tooth number step;
In the information processing device, it is determined whether it is a tooth or an implant for each tooth number by using the second image classification model learned to distinguish between a tooth and an implant and generate information on a state of a tooth and an implant, and when it is determined as a tooth, the information processing device determines whether the tooth is an implant. Generating treatment information of the tooth, and generating implant information related to the implant when it is determined to be an implant; and
In the information processing device, generating a tooth state interpretation statement by combining the information on whether the tooth is a tooth or an implant for each tooth number, the treatment information, and the implant information;
Calculating the location information for the tooth,
In the information processing device, distinguishing individual teeth from the panoramic tooth image using the first object detection module; and
In the information processing device, calculating a first vertical maximum value and a first vertical minimum value for each of the separated teeth as coordinates,
Calculating the location information of the alveolar line,
calculating alveolar lines of the teeth in the panoramic tooth image using the second object detection module in the information processing device; and
In the information processing device, the second vertical maximum value is calculated as coordinates using upper alveolar lines among the calculated alveolar lines, and the second vertical minimum value is calculated as coordinates using lower alveolar lines among the calculated alveolar lines. Including the step of calculating as,
The step of distinguishing whether the teeth are upper or lower teeth,
In the information processing device, comparing the first vertical maximum value, the first vertical minimum value, the second vertical maximum value, and the second vertical minimum value to distinguish whether the tooth is an upper jaw or a lower jaw. A method for generating a dental condition reading statement using artificial intelligence. The method of claim 8, wherein the treatment information,
Including at least one of information on whether or not the tooth is treated and information on whether or not to treat the nerve,
The implant information,
A method of generating a dental condition reading statement using artificial intelligence, characterized in that it includes at least one of manufacturer information of the implant and model information of the implant. delete The method of claim 8, wherein the step of distinguishing whether the teeth are upper or lower teeth,
determining, in the information processing device, that the tooth is an upper jaw when the first vertical maximum value is greater than or equal to the second vertical maximum value and the first vertical minimum value is greater than or equal to the second vertical minimum value; and
In the information processing device, if the first vertical minimum value is less than or equal to the second vertical minimum value and the first vertical maximum value is less than or equal to the second vertical maximum value, determining the tooth as a mandibular tooth. A method for generating a dental condition reading statement using intelligence. The method of claim 8, wherein the step of distinguishing whether the teeth are upper or lower teeth,
In the information processing device, determining that the tooth is an upper jaw when the second vertical maximum value is between the first vertical maximum value and the first vertical minimum value or when the first vertical minimum value is greater than or equal to the second vertical maximum value ; and
In the information processing device, determining that the tooth is a mandibular tooth when the second vertical minimum value is between the first vertical maximum value and the first vertical minimum value or when the first vertical maximum value is equal to or less than the second vertical minimum value A method of generating a dental condition reading statement using artificial intelligence, characterized in that it comprises. The method of claim 8, wherein the step of distinguishing whether the teeth are upper or lower teeth,
In the information processing device, when the first vertical maximum value is less than the second vertical maximum value and the first vertical minimum value exceeds the second vertical minimum value, a distance between the first vertical minimum value and the second vertical maximum value If the distance is smaller than the distance between the first vertical maximum value and the second vertical minimum value, it is determined as an upper jaw, and the distance between the first vertical minimum value and the second vertical maximum value is the first vertical maximum value and the second vertical minimum value. 2. A method for generating a dental condition reading statement using artificial intelligence, further comprising the step of determining that the lower teeth are greater than the distance between the vertical minimum values. The method of claim 8, wherein the step of assigning the tooth number,
generating, in the information processing device, upper jaw tooth groups in all cases to which tooth numbers can be applied according to the number of teeth determined as the upper jaw teeth;
calculating, in the information processing device, a probability that each tooth matches a tooth formula by using the first image classification model for each tooth group of the upper teeth;
assigning, in the information processing device, tooth numbers of teeth corresponding to the maxillary teeth as the tooth numbers of the maxillary tooth groups selected using the score calculated using the calculated probability for each of the maxillary tooth groups;
generating, in the information processing device, tooth groups of lower jaw teeth in all cases to which tooth numbers can be applied according to the number of teeth determined as the lower jaw teeth;
calculating, in the information processing device, a probability that each tooth matches a tooth formula by using the first image classification model for each tooth group of the lower jaw; and
In the information processing device, assigning tooth numbers of teeth corresponding to the lower jaw teeth as the tooth numbers of the lower jaw tooth groups selected using the score calculated using the calculated probability for each lower jaw tooth group. Method for generating a tooth condition reading statement using artificial intelligence, characterized in that. According to claim 14,
The scores for each tooth group of the teeth corresponding to the upper or lower teeth are,

(N is the number of teeth of the upper jaw tooth group or the lower jaw tooth group,

is the probability that the ith tooth (ti) is the ith tooth formula (si)
Method for generating a dental condition reading statement using artificial intelligence, characterized in that calculated as.

Images