Image processing device and method, program, and storage medium

Abstract

To provide an image processing device that can obtain results that are consistent with past medical information when inferring the type and condition of teeth from intraoral images. [Solution] The system comprises a first acquisition unit that acquires the patient's past medical information, a second acquisition unit that acquires intraoral images of the patient, an inference unit that infers the tooth type from the intraoral images, a correction unit that corrects the inference results by the inference unit based on predetermined rules to make them consistent with the past medical information, and an output unit that outputs the tooth type information corrected by the correction unit.

Description

[Technical Field]

[0001] The present invention relates to an image processing device that detects tooth types from intraoral images. [Background technology]

[0002] An image processing device is known that infers the type and condition of teeth from intraoral photographs based on machine learning (see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-078940 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in cases where there are congenitally missing teeth or when teeth have been extracted, tooth numbers are sometimes determined by a rule of convenience rather than by the biological tooth type. For example, if an incisor is missing and it is unclear whether number one or number two is missing, number two will be considered missing. If it is unclear whether number four or number five premolar is missing, number four will be considered missing. If one of number six, seven, or eight molars is missing and it is unclear which is missing, the larger number will be considered missing.

[0005] In the technology of Patent Document 1, tooth type and tooth condition are inferred from intraoral images, so the tooth numbers determined by inference may differ from tooth numbers determined for convenience in past medical treatment, resulting in inconsistencies with past medical information.

[0006] The present invention has been made in consideration of the above-mentioned problems, and its purpose is to provide an image processing device that can obtain results that are consistent with past medical information when inferring the type and condition of teeth from intraoral images. [Means for solving the problem]

[0007] The image processing device of the present invention is characterized by comprising a first acquisition means for acquiring past medical information of a target patient, a second acquisition means for acquiring intraoral images of the patient, an inference means for inferring tooth type from the intraoral images, a correction means for correcting the result of inference by the inference means based on predetermined rules so that it is consistent with the past medical information, and an output means for outputting information on tooth type corrected by the correction means. [Effects of the Invention]

[0008] According to the present invention, when inferring the type and condition of teeth from intraoral images, it is possible to obtain results that are consistent with past medical information. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a block diagram showing the configuration of a dental system according to a first embodiment. [Figure 2] FIG. 1 is a hardware configuration diagram of an image processing apparatus. [Figure 3] 10 is a flowchart showing a process from reading an oral cavity image to notification. [Figure 4] A conceptual diagram showing the inference results for tooth number and condition. [Figure 5] A diagram showing the correspondence between tooth classification and tooth number. [Figure 6] 10 is a flowchart showing a process of associating tooth numbers with states. [Figure 7] 10 is a flowchart showing a correction process for a permanent dentition. [Figure 8] A diagram showing the correspondence between past medical information, inference results, and corrected results. [Figure 9] 10 is a flowchart showing the process of erupting a partially impacted tooth. [Figure 10A] 10 is a flowchart showing a correction process for mixed dentition. [Figure 10B] 10 is a flowchart showing a correction process for mixed dentition. [Figure 11]A diagram showing the relationship between age and the presence or absence of teeth. [Figure 12] A diagram showing the correspondence of teeth that should not coexist. [Figure 13] A diagram showing the correspondence between past medical information, inference results, and corrected results. [Figure 14] Conceptual diagram showing past medical information. [Figure 15] A diagram showing the correspondence between past medical information, inference results, and corrected results. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the invention claimed. Although multiple features are described in the embodiments, not all of these multiple features are necessarily essential to the invention, and multiple features may be combined arbitrarily. Furthermore, in the accompanying drawings, the same reference numerals are used to designate the same or similar components, and redundant explanations will be omitted.

[0011] (First embodiment) In this embodiment, when inferring the type and condition of teeth based on an intraoral image of the mandibular occlusal view of an adult with only permanent teeth present, a method is described for correcting the inference results so that they are consistent with past medical information based on the specified rules described below.

[0012] [Definition of words] Intraoral photographs are visible light image data taken using the five-photo method used in dental diagnosis. The five images in the five-photo method consist of a front view, an occlusal view of the upper jaw, an occlusal view of the lower jaw, a left lateral view, and a right lateral view.

[0013] In this embodiment, the tooth number is a combination of letters and numbers that uniquely identifies the position and type of tooth in the oral cavity. For teeth in the upper jaw, the first character is u, and for teeth in the lower jaw, the first character is b. For teeth on the right side of the midline, the second character is R, and for teeth on the left side, it is L. The third character is a number assigned in ascending order from the front to the back. For permanent teeth, central incisors are numbered 1, and third molars are numbered 8. For baby teeth, baby central incisors are numbered A, and second baby molars are numbered E.

[0014] Based on this convention, teeth located on the right side of the patient's maxilla will be labeled uR1 through uR8, numbered 1 through 8. Teeth located on the left side of the patient's maxilla will be labeled uL1 through uL8, numbered 1 through 8. Teeth located on the right side of the patient's mandible will be labeled bR1 through bR8, numbered 1 through 8. Teeth located on the left side of the patient's mandible will be labeled bL1 through bL8, numbered 1 through 8.

[0015] For primary teeth, the teeth located on the right side of the patient's maxilla will be labeled A through E, uRA through uRE. The teeth located on the left side of the patient's maxilla will be labeled A through E, uLA through uLE. The teeth located on the right side of the patient's mandible will be labeled A through E, bRA through bRE. The teeth located on the left side of the patient's mandible will be labeled A through E, bLA through bLE.

[0016] In contrast, tooth type is sometimes used as a name based on the shape of the tooth. From the front on each of the top, bottom, left and right, they are central incisors, lateral incisors, canines, first premolars, second premolars, first molars, second molars, and third molars. In the case of primary teeth, they are primary central incisors, primary lateral incisors, primary canines, first primary molars, and second primary molars. When referring to tooth types, it can be unclear whether they are on the upper or lower jaw or the left or right side, so in this embodiment, tooth numbers are used to identify individual teeth on the upper and lower jaws.

[0017] Tooth classification is a group of teeth with similar shapes, and refers to incisors, canines, premolars, molars, primary incisors, primary canines, and primary molars. The central incisors and lateral incisors are called incisors, the first premolars and second premolars are called premolars, and the first molars, second molars, and third molars are called molars. In primary teeth, the primary central incisors and lateral incisors are called primary incisors, and the first molars and second molars are called primary molars.

[0018] The tooth status includes healthy teeth, metal partial restorations, white partial restorations, full metal crowns, cavities, etc., and also includes artificial teeth such as pontics, implants, and dentures. Since artificial teeth are made to resemble real teeth, it is possible to infer that a tooth exists as healthy even when there is no actual tooth. In many cases, only part of a tooth is visible during eruptions, in which case it is inferred as being partially impacted. In such cases, it may not be possible to infer the type of tooth.

[0019] The tooth type information is assumed to be the type of tooth that can be identified by the tooth number and whether or not it exists.

[0020] The treatment information is assumed to be information relating to treatment such as partial filling, replacement with an artificial tooth, orthodontics, and tooth extraction.

[0021] The medical information includes the patient's name, sex, age, medical date and time, tooth type information, and treatment information.

[0022] In this embodiment, machine learning is used to infer tooth type and tooth condition. An intraoral image is input, and a learning model is used that learns the corresponding actual tooth type and tooth condition as training data to infer tooth type and tooth condition from the intraoral image. Note that any image processing that can produce similar processing results may be used, such as pattern matching processing, edge detection, or discrimination processing using shading or color information.

[0023] [System Configuration] 1 is a block diagram showing the device configuration of a dental system 100 according to this embodiment. The dental system 100 according to this embodiment is configured to include devices that perform the following processes.

[0024] The image processing device 101 of this embodiment determines each dental formula number and tooth condition, and then corrects the results using a method described below.

[0025] The imaging device 108 captures an image of the patient's oral cavity 109 and generates visible light image data. In the following explanation, the image captured by the imaging device 108 will be described as visible light image data. Furthermore, patient information such as a patient ID associated with the patient is added to the captured image data and transmitted to the image processing device 101. Although not shown, the imaging device 108 is assumed to be equipped with an input / output device for checking the captured image and selecting patient information, similar to a commercially available digital camera.

[0026] The dental electronic medical record system 104 receives a request from the dental electronic medical record display terminal 103 (described later), communicates with the image processing device 101 and the medical information DB 105 (described later), and transmits and receives data associated with the dental electronic medical record creation process.

[0027] The dental electronic medical record display terminal 103 communicates with the dental electronic medical record system 104, and displays the dental electronic medical record and accepts input when creating it.

[0028] The medical information DB 105 communicates with the dental electronic medical record system 104 and stores, transmits and receives the patient's medical information so as to associate it with patient information such as the patient ID.

[0029] The in-hospital system 106 communicates with a patient information DB 107 (described later) to register and acquire patient information, and also communicates with the dental electronic medical record system 104 to transmit patient information.

[0030] The patient information DB 107 communicates with the in-hospital system 106 and stores, transmits, and receives patient information.

[0031] The patient's oral cavity 109 represents the object of examination in the dental system.

[0032] The image processing device 101, imaging device 108, electronic medical record system 104, and in-hospital system 106 communicate via a network 102. Furthermore, communication between the dental electronic medical record display terminal 103 and the dental electronic medical record system 104 is performed using an HDMI (registered trademark) cable or the like. Communication between the medical information DB 105 and the dental electronic medical record system 104, and communication between the in-hospital system 106 and the patient information DB 107 is performed using a USB cable or the like. The communication means are, by way of example, a network, HDMI (registered trademark), and USB, but are not limited to these.

[0033] The dental electronic medical record system 104 acquires information about the patient to be examined from the patient information DB 107 via the in-hospital system 106 .

[0034] The image processing device 101 also communicates with the imaging device 108 and the dental electronic medical record system 104, and acquires patient information from the dental electronic medical record system 104 that is the subject of the examination when the imaging device 108 is turned on or the patient information is reconfirmed. The image processing device 101 then transmits the patient information to the imaging device 108. The image processing device 101 receives intraoral images captured by the imaging device 108 via the network 102. The intraoral images are assumed to be associated with patient information, such as information on the lower jaw, upper jaw, entire surface, left side, right side, and whether the image was captured using a mirror.

[0035] The image processing device 101 infers the tooth numbers and conditions from the intraoral images received from the imaging device 108 and site information such as the upper and lower jaws. At that time, based on the patient information associated with the intraoral images, the image processing device 101 acquires the patient's past medical information from the medical information DB 105 via the electronic medical record system 104 and corrects the inference results using a method described below. The corrected tooth numbers and conditions are sent from the image processing device 101 to the dental electronic medical record system 104 and displayed on the dental electronic medical record display terminal 103.

[0036] [Configuration of image processing device] FIG. 2 is a diagram showing the hardware configuration of the image processing apparatus 101 according to this embodiment.

[0037] The image processing device 101 has a CPU 201, which is a processor that processes programs, and a ROM (Read Only Memory) 202, which is memory that stores the programs. It also has a RAM (Random Access Memory) 203, which is memory that expands data required when the programs are executed. It also has an HDD (Hard Disk Drive) 205, which is a hard disk that stores inference data, inference results, and data for generating inference data, such as learning data. It also includes an input device 206 used when registering setting information for the programs, a display 204 for display confirmation, an I / F (Interface) 207 for communicating with external systems, a bus 208, etc.

[0038] Each function of the image processing device 101 is realized by loading a predetermined program onto hardware such as a CPU 201 and a ROM 202 and having the CPU 201 perform calculations. Each function is also realized by controlling communication with an external system using an I / F 207 and reading and writing data from and to a RAM 203 and a hard disk drive (HDD) 205.

[0039] When an image is written from the imaging device 108 to the RAM 203 via the I / F 207, the CPU 201 performs vertical or horizontal flip processing on the image as appropriate based on the associated information, and then stores the image in the HDD 205.

[0040] [Change notification processing] Fig. 3 is a flowchart showing the processing from image reading to notification in the image processing device 101. A series of processing from inputting an image to be analyzed to notifying medical personnel of changes that require attention will be described using Fig. 3. An image of the occlusal surface of the mandible will be described as the subject.

[0041] In step S301, the CPU 201 reads the intraoral image to be analyzed from the HDD 205 into the RAM 203.

[0042] In step S302, the CPU 201 infers the tooth number from the image. Figure 4(a) shows an example of the inferred tooth number. Figure 4 will be described later.

[0043] In step S303, the CPU 201 infers the tooth condition. Figure 4(b) shows an example of the inference result of the tooth condition.

[0044] In step S304, the CPU 201 associates the tooth number with the state. The method of associating the tooth number with the state will be described later with reference to FIG.

[0045] In step S305, the CPU 201 reads past medical information of the patient corresponding to the image read in step S301 into the RAM 203. The past medical information of the patient is obtained from the medical information DB 105 via the dental electronic medical record system 104 based on the patient information associated with the image. The past medical information of the patient is then stored in the RAM 203.

[0046] In step S306, the CPU 201 corrects the inference result based on past medical information. The correction method will be described later.

[0047] In step S307, the CPU 201 outputs the tooth numbers and states corrected in step S306. Examples of the output results will be described later. The image processing device 101 transmits this to the dental electronic medical record system 104 as dentition information in the electronic medical record of the corresponding patient. The user can check the results on the dental electronic medical record display terminal 103.

[0048] In step S308, the CPU 201 notifies the user of any changes that require attention. The notification is made by marking the relevant tooth on the electronic medical record screen, or by displaying a speech bubble on the relevant tooth if the notification content can be expressed in words. If a tooth that was previously healthy has now developed a cavity, the user is notified that a cavity is suspected. If a tooth that was previously missing is detected, the user is notified that it may be an artificial tooth.

[0049] [Example of inference results] 4A and 4B are conceptual diagrams showing inference results in this embodiment. Fig. 4A shows an example of the inference results for tooth numbers, and Fig. 4B shows an example of the inference results for tooth conditions. For the purpose of explaining the correction process described below, the inference results are assumed to be the results shown in 401 to 412 and 414 to 419 below.

[0050] Reference numeral 400 denotes an image to be processed for tooth number inference. This is an intraoral image of the mandibular occlusion of an adult with only permanent teeth present.

[0051] 401 to 412 are rectangles resulting from detecting tooth numbers using a machine learning learning model, and each rectangle is associated with a label of the corresponding tooth number. Each rectangle is also associated with coordinate information on the image. The CPU 201 executes inference, and the result is stored in the HDD 205 in association with the image to be inferred.

[0052] 401 is inferred to be a third molar on the right side of the mandible and is associated with the label bR8.

[0053] 402 is inferred to be the first molar on the right side of the mandible and is associated with the label bR6.

[0054] 403 is inferred to be the second premolar on the right side of the mandible and is associated with the label bR5.

[0055] 404 is inferred to be the first premolar on the right side of the mandible and is associated with the label bR4.

[0056] 405 is inferred to be a canine tooth on the right side of the mandible and is associated with the label bR3.

[0057] 406 is inferred to be the right lower lateral incisor and is associated with the label bR2.

[0058] 407 is inferred to be the left mandibular central incisor and is associated with the label bL1.

[0059] 408 is inferred to be the left mandibular lateral incisor and is associated with the label bL2.

[0060] 409 is inferred to be the left mandibular canine and is associated with the label bL3.

[0061] 410 is inferred to be the first premolar on the left side of the mandible and is associated with the label bL4.

[0062] 411 is inferred to be the left mandibular first molar and is associated with the label bL6.

[0063] 412 is inferred to be the left mandibular second molar and is associated with the label bL7.

[0064] Reference numeral 413 denotes an image to be processed for inferring the state of the teeth. It is assumed to be the same image as 400.

[0065] 414 to 419 are rectangles resulting from detecting tooth conditions using a machine learning learning model, and each rectangle is associated with a label of the corresponding tooth condition.

[0066] 414 is associated with the label Metallic Partial Repair.

[0067] The 415 is associated with the label Full Metal Crown (hereafter FMC).

[0068] 416 is associated with the label of acid erosion.

[0069] 417 has the label associated with acid erosion.

[0070] 418 is associated with the label of acid erosion.

[0071] 419 has the associated label of tooth decay.

[0072] [Tooth Classification] 5 is a correspondence table between tooth classifications and tooth numbers in this embodiment, which will be used in the correction process described later.

[0073] 501 is a column of a category ID that uniquely identifies a category.

[0074] 502 is a tooth classification column indicating the names of each tooth classification.

[0075] 503 is a string of tooth numbers included in each category.

[0076] Lines 504 to 510 are related to permanent teeth, and lines 511 to 515 are related to baby teeth.

[0077] 504 is a line showing the incisors and the tooth numbers included in that classification.

[0078] 505 is a line showing the left canine and the tooth numbers included in that classification.

[0079] 506 is a line showing the left premolar and the tooth numbers included in that classification.

[0080] 507 is a line showing the left molars and the tooth numbers included in that classification.

[0081] 508 is a line showing the right canine and the tooth numbers included in that classification.

[0082] 509 is a line showing the right premolar and the tooth numbers included in that classification.

[0083] 510 is a row showing the right molars and the tooth numbers included in that classification.

[0084] 511 is a line showing the numbers of the deciduous incisors and the teeth included in that classification.

[0085] 512 is a line showing the left deciduous canine and the tooth numbers included in that classification.

[0086] 513 is a line showing the left deciduous molars and the tooth numbers included in that classification.

[0087] 514 is a line showing the right deciduous canine and the tooth numbers included in that classification.

[0088] 515 is a line showing the right deciduous molars and the tooth numbers included in that classification.

[0089] [How to associate tooth numbers with conditions] 6 is a flowchart showing the process of associating tooth numbers and states in the image processing device 101. The process of associating tooth numbers and states in the inference results shown in FIG. 4 will be described with reference to FIG.

[0090] In step S601, the CPU 201 reads the inference results of the tooth number and state from the HDD 205 into the RAM 203.

[0091] In step S602, the CPU 201 assigns 0 to a counter cnt for executing a series of processes for the inferred state of each tooth.

[0092] In step S603, the CPU 201 determines whether cnt is smaller than the number of inferred states. If it is smaller, the CPU 201 proceeds to step S604. If Cnt is equal to or greater than the number of inferred states, the CPU 201 determines that processing has been completed for all tooth states and ends the series of processes.

[0093] In step S604, the CPU 201 reads the coordinates of the rectangle in the cnt-th state. The origin of the coordinates in the image is assumed to be the upper left, the X-axis coordinate increases horizontally to the right, and the Y-axis coordinate increases vertically downward. The inferred rectangles are then processed in order starting with the rectangle with the smallest X-coordinate of the top left vertex. If there are multiple rectangles with the same X-coordinate, the rectangle with the smallest Y-coordinate is processed first.

[0094] In step S605, the CPU 201 searches for a rectangle of a tooth number that overlaps with the rectangle of the cnt-th state by 80% or more in area.

[0095] In step S606, the CPU 201 determines whether a rectangle with a tooth number that overlaps with the rectangle of the cnt-th state by 80% or more in area was found in step S605. If found, the CPU 201 proceeds to step S607. If not found, the CPU 201 proceeds to step S608.

[0096] In step S607, the CPU 201 stores the cnt-th state in the HDD 205 in association with the tooth number, assuming that the state is the state of the tooth number where 80% or more of the area overlaps.

[0097] In step S608, the CPU 201 stores the cnt-th state as not associated with a tooth, without associating it with a tooth number, in the HDD 205. States without a tooth shape, such as impacted teeth and residual roots, fall into this branch.

[0098] In step S609, the CPU 201 increments the counter cnt, and then returns the process to step S603.

[0099] [Correction processing] Fig. 7 is a flowchart showing the correction process for the permanent dentition in the image processing device 101. A method for correcting the inference result using past medical information and ensuring consistency with the past medical information will be described with reference to Fig. 7.

[0100] In step S700, the CPU 201 classifies the tooth numbers present in the past medical treatment information stored in the RAM 203 into the categories shown in FIG.

[0101] In step S701, the CPU 201 classifies the tooth numbers of the inference results stored in the RAM 203 into the categories shown in FIG.

[0102] In step S702, the CPU 201 substitutes the number of classifications that have one or more classified teeth based on the past medical information and the inference results into a value max used to determine the termination condition.

[0103] In step S703, the CPU 201 assigns 1 to a counter n for executing processing for each classification.

[0104] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. If the counter n is equal to or less than max, the CPU 201 proceeds to step S705. If the counter n is greater than max, the CPU 201 determines that the correction process executed for each classification has ended, and proceeds to step S717.

[0105] In step S705, the CPU 201 determines whether the number of teeth classified into the n-th tooth category 502 matches the past medical information and the inference result. If they match, the CPU 201 proceeds to step S706. If they do not match, the CPU 201 proceeds to step S709.

[0106] In step S706, the CPU 201 determines whether the tooth in the past medical information included in the n-th tooth classification 502 and the tooth in the inference result have the same tooth number. If they are the same, the CPU 201 proceeds to step S714. If they are not, the CPU 201 proceeds to step S707.

[0107] In step S707, the CPU 201 changes the tooth numbers in the inference results based on the coordinates so that they match the past medical results. For the incisor 504, the CPU 201 corrects the coordinates so that the center coordinates of the rectangles representing the tooth numbers are in ascending order of bR2, bR1, bL1, and bL2 along the X axis, and so that they match the past medical results. For the left molar 507, the CPU 201 corrects the coordinates so that the center coordinates of the rectangles representing the tooth numbers are in ascending order of bL8, bL7, and bL6 along the Y axis, and for the right molar 510, the CPU 201 corrects the coordinates so that the center coordinates of the rectangles representing the tooth numbers are in ascending order of bR8, bR7, and bR6 along the Y axis, and so that they match the past medical results.

[0108] In step S708, the CPU 201 associates the corrected tooth number and state of the inference result classified into the n-th tooth classification 502, and stores the association result as corrected information.

[0109] In step S709, the CPU 201 determines whether the number of teeth classified into the n-th tooth category 502 has increased in the inference result compared to the past medical information. If the number has increased, the CPU 201 proceeds to step S710. If the number has decreased, the CPU 201 proceeds to step S712.

[0110] In step S710, the CPU 201 determines which teeth have been added after verifying consistency between the past medical records and the inference results. The method of verifying consistency is the same as in step S707. If there are any missing teeth in the past medical records, the positions of those missing teeth will correspond to the added teeth. If there are multiple missing teeth and the added teeth are fewer than the missing teeth, the missing teeth will be determined according to the expedient rules described above.

[0111] In step S711, the CPU 201 determines that a tooth added without a treatment record in the past medical results may be an artificial tooth such as an implant. In addition, if the past medical records do not show that the tooth is an artificial tooth, the tooth number is associated with the tooth and the tooth is made the target of notification. In step S712, the CPU 201 compares the past medical examination results with the inference results, and extracts the numbers of the teeth that have been removed (decreased) after ensuring consistency between the past medical examination results and the inference results. The method of ensuring consistency is the same as in step S707.

[0112] In step S713, the CPU 201 determines that teeth that have been lost without treatment records in past medical examination results may have been lost due to extraction, fracture, periodontal disease, etc. Therefore, the CPU 201 associates the teeth with tooth numbers and makes them the subject of notification. In step S714, the CPU 201 stores the tooth number and state of the inference result classified into the n-th tooth classification 502 as corrected.

[0113] In step S715, the CPU 201 stores the tooth state of the past medical treatment results classified in the nth tooth category 502 as corrected information if there is no contradiction with the inference result or the result of determining that the tooth is an artificial tooth. Note that the tooth state of the inference result has already been stored in step S708. If the past medical treatment result is an artificial tooth such as an implant, pontic, or bridge, and the inference result is a sound tooth, there is no contradiction, and the state of the past medical treatment result is stored as a corrected state. If the past medical treatment result is a full metal crown or partial restoration, and the inference result is a sound tooth, there is a contradiction and this is subject to notification.

[0114] In step S716, the CPU 201 increments the counter n for executing processing for each classification, and returns the process to step S704.

[0115] In step S717, the CPU 201 refers to the corrected tooth number and state, and if there is a disease in the state, the tooth number is set as the notification target, and the series of correction processes is then terminated.

[0116] [Example before and after correction] Figure 8 shows an example of the correspondence between past medical information, inference results, and corrected results. Figure 8 corresponds to the inference results in Figure 4.

[0117] 801 is the tooth number sequence.

[0118] 802 is a past record column that holds past medical treatment results. A circle indicates that the tooth was present, and an x ​​indicates that the tooth was not present. In addition, in the case of a healthy tooth, the status is not particularly displayed, and if there is treatment information such as an artificial tooth or partial restoration, that status is stored. 803 is an inference result string that holds the tooth number and tooth status of the inference result. ◯ indicates that the tooth was inferred to exist, and × indicates that the tooth was not found by inference. In addition, if the tooth is healthy, the status is not displayed specifically, but if there is a disease or traces of treatment that can be inferred as the status, that status is stored.

[0119] Reference numeral 804 denotes a post-correction column that holds the presence or absence and status of teeth after correction. ◯ indicates that a tooth has been determined to be present after correction, and × indicates that a tooth has been determined to be absent. × also indicates that a tooth is missing. In addition, in the case of a healthy tooth, the status is not displayed, and if it is determined to be a diseased or treatment scar or an artificial tooth, that status is stored. The more elaborate the artificial teeth such as pontics, implants, and dentures are, the more they appear to be healthy teeth, but since they are not actual teeth, doctors will determine that they are missing. The image processing device 101 of this embodiment also outputs them as missing.

[0120] 805 is a notification column. If any unusual changes or diseases that require attention are found, a notification is made. If no changes that require particular attention are found, no notification is made.

[0121] 806 to 821 are rows that hold information about each tooth.

[0122] 806 is a row that holds information about bL1, 807 is bL2, 808 is bL3, 809 is bL4, 810 is bL5, 811 is bL6, 812 is bL7, and 813 is a row that holds information about bL8.

[0123] 814 is a row that holds information about bR1, 815 is bR2, 816 is bR3, 817 is bR4, 818 is bR5, 819 is bR6, 820 is bR7, and 821 is a row that holds information about bR8.

[0124] The results of the correction process will be described with reference to Figures 7 and 8. In this embodiment, the description will be made on the lower jaw teeth of an adult who has only permanent teeth.

[0125] In step S700, the CPU 201 classifies the tooth numbers present in the past medical information stored in the RAM 203 into the categories explained in Fig. 5. When the information in the past record sequence 802 is divided, the incisor 504 is bL1, bL2, and bR1, the left canine 505 is bL3, the left premolar 506 is bL5, the left molar 507 is bL7, the right canine 508 is bR3, the right premolar 509 is bR5, the right molar 510 is bR6, bR7, and bR8, and there are no baby molars.

[0126] In step S701, the CPU 201 classifies the tooth numbers of the inference results stored in the RAM 203 into the categories explained in Fig. 5. When the information in the inference result column 803 is divided, the incisor 504 is bL1, bL2, and bR2, the left canine 505 is bL3, the left premolar 506 is bL4, the left molar 507 is bL6 and bL7, the right canine 508 is bR3, the right premolar 509 is bR4 and bR5, the right molar 510 is bR6 and bR8, and there are no baby molars.

[0127] In step S702, the CPU 201 assigns the number of categories that have one or more teeth classified based on the past medical information and the inference results to the value max used to determine the termination condition. Since there is one or more teeth in each category, from the incisor 504 to the right molar 510, max is 7.

[0128] In step S703, the CPU 201 assigns 1 to a counter n for executing processing for each classification.

[0129] [Incisor] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 1, it is determined that the counter n is equal to or less than max, and the process proceeds to step S705.

[0130] In step S705, CPU 201 determines whether the number of teeth classified into first tooth category 502 matches the past medical information and the inference result. Since the number of teeth in the past medical information included in incisor 504 is three and the inference result is also three, it is determined that they match, and the process proceeds to step S706.

[0131] In step S706, the CPU 201 determines whether the teeth in the past medical information included in the incisor 504 are the same as the teeth in the inference result. In the past medical information, bL1, bL2, and bR1 are included, whereas the inference result is bL1, bL2, and bR2. Therefore, it is determined that they do not match, and the process proceeds to step S707.

[0132] In step S707, the CPU 201 changes the tooth type in the inference result based on the coordinates to match the past medical treatment results. In the past medical treatment information, bR1 is a remaining tooth, while the inference result is bR2, so bR2 in the inference result is changed to bR1. In step S708, the CPU 201 stores the corrected tooth number and state of the inference result classified as incisor 504 as corrected information. Since bR2 was associated with acid erosion in the inferred tooth state, bR1 is set to acid erosion. The result is as shown in row 814 of bR1 in the correction column 804. In addition, since bR2 in the inference result has been changed to bR1, bR2 becomes missing. The result is as shown in row 815 of bR2 in the correction column 804. Since bL1 and bL2 did not require correction, the inference result state of acid erosion is associated with each tooth and stored. This is as shown in row 806 of bL1 and row 807 of bL2 in the correction column 804.

[0133] In step S715, the CPU 201 stores the tooth condition in the past medical records classified as incisor 504 as corrected information if there is no contradiction with the inference result or the result of determining that the tooth is an artificial tooth. In the past medical records, bL1, bL2, and bR1 are all sound teeth, and the inference result is that they are all acid erosion, so there is no contradiction. However, because there is no need to record the state of sound teeth, the tooth condition in the past medical records is not recorded.

[0134] In step S716, the CPU 201 increments a counter n for executing processing for each classification. Counter n becomes 2.

[0135] [Left canine] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 2, the CPU 201 determines that the counter n is equal to or less than max, and proceeds to step S705.

[0136] In step S705, CPU 201 determines whether the number of teeth classified into second tooth category 502 matches the past medical information and the inference result. Since the number of teeth included in left canine 505 in the past medical information is one and the inference result is also one, it is determined that they match, and the process proceeds to step S706.

[0137] In step S706, the CPU 201 determines whether the tooth in the past medical information included in the left canine 505 is the same as the tooth in the inference result. Since both the past medical information and the inference result are bL3 and match, the process proceeds to step S714.

[0138] In step S714, the CPU 201 stores the tooth number and state of the inference result classified as the left canine 505 as corrected, as shown in row 808 of bL3 in the correction column 804.

[0139] In step S715, the CPU 201 stores the state of the tooth classified as left canine 505 in the past medical records as corrected information if there is no contradiction with the inference result or the result of determining that it is an artificial tooth. bL3 is a sound tooth in both the past medical records and the inference result, so there is no contradiction. However, because there is no need to record the state of being a sound tooth, the state of the tooth in the past medical records is not recorded.

[0140] In step S716, the CPU 201 increments a counter n for executing processing for each classification.

[0141] [Left premolar] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 3, it is determined that the counter n is equal to or less than max, and the process proceeds to step S705.

[0142] In step S705, CPU 201 determines whether the number of teeth classified into third tooth category 502 matches the past medical information and the inference result. Since the number of teeth included in left premolar 506 in the past medical information is one and the inference result is also one, it is determined that they match and the process proceeds to step S706.

[0143] In step S706, CPU 201 determines whether the tooth in the past medical information included in left premolar 505 is the same as the tooth in the inference result. Since the past medical information is bL5 and the inference result is bL4, it is determined that they do not match, and the process proceeds to step S707.

[0144] In step S707, the CPU 201 changes the tooth type of the inferred result to match the past medical result based on the coordinates. In the past medical information, bL5 is the remaining tooth, while the inferred result is bL4, so bL4 in the inferred result is changed to bL5.

[0145] In step S708, the CPU 201 associates the tooth status inferred with the tooth number corrected in step S707. The inferred tooth status is that there is no status associated with bL4, i.e., it is a healthy tooth, so bL5 also has no status. Note that because bL4 in the inference result was changed to bL5, bL4 is missing. This is as shown in row 809 for bL4 and row 810 for bL5 in the correction column 804.

[0146] In step S715, the CPU 201 stores the state of the tooth in the past medical record classified as left premolar 505 as corrected information if there is no contradiction with the inference result or the result of determining that it is an artificial tooth. bL4 is missing in the past medical record and the state associated in step S708 is also missing, so there is no contradiction. bL5 is a sound tooth in the past medical record and the inference result associated in step S708 is also a sound tooth, so there is no contradiction. However, because there is no need to record the state of being a sound tooth, the state of the tooth in the past medical record is not recorded.

[0147] In step S716, the CPU 201 increments a counter n for executing processing for each classification.

[0148] [Left molar] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 4, it is determined to be equal to or less than max, and the process proceeds to step S705.

[0149] In step S705, CPU 201 determines whether the number of teeth classified into fourth tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in left molar 507 in the past medical information is one, and the inference result is two, it is determined that there is no match, and the process proceeds to step S709.

[0150] In step S709, CPU 201 determines whether the number of teeth classified as left molar 507 has increased in the inference result compared to the past medical information. Since the number of teeth included in left molar 507 in the past medical information is one, but the inference result is two, it is determined that the number has increased, and the process proceeds to step S710.

[0151] In step S710, CPU 201 checks the consistency of the past medical records with the inference results and then determines which teeth have been added. The tooth number in the past medical information included in left molar 507 is only bL7, whereas the tooth numbers in the inference results are bL6 and bL7. In the past medical records, bL6 is missing, and bL7 in the inference results is also associated with a cavity. Therefore, bL6 is determined to be the tooth that has been added.

[0152] In step S711, the CPU 201 determines that a tooth added without a past treatment record may be an artificial tooth such as an implant. In addition, if the past medical records do not indicate that the tooth is an artificial tooth, the tooth number is associated with the tooth and the tooth is subject to notification. Row 811 of bL6 in the past record column 802 is missing, so the tooth is subject to notification.

[0153] In step S708, CPU 201 stores the tooth number and state of the inference result classified as left molar 507 as corrected. Because bL6 was inferred to be a sound tooth despite being missing, it is stored as the state of an artificial tooth, as shown in row 811 of bL6 in correction column 804. The state of bL7 is stored as a carious tooth, as shown in row 812 of bL7 in correction column 804.

[0154] In step S715, CPU 201 stores the tooth status in the past medical records classified as left molar 507 as corrected if it is consistent with the inference result or the result of determining that it is an artificial tooth. bL6 is also missing in the past medical records and has been determined to be an artificial tooth in step S711, so row 811 for bL6 in correction column 804 is not changed. bL7 is a sound tooth in the past medical records and a cavity in the inference result, so there is no contradiction. However, since there is no need to record the status of a sound tooth, the tooth status in the past medical records is not recorded. bL8 is missing in both the past medical records and the inference result, so there is no contradiction. Therefore, the state of the past medical treatment is stored. The result is as shown in row 813 for bL8 in correction column 804.

[0155] In step S716, the CPU 201 increments a counter n for executing processing for each classification.

[0156] [Right canine] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 5, it is determined to be equal to or less than max, and the process proceeds to step S705.

[0157] In step S705, CPU 201 determines whether the number of teeth classified into second tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in right canine 508 in the past medical information is one and the inference result is also one, it is determined that they match, and the process proceeds to step S706.

[0158] In step S706, the CPU 201 determines whether the tooth in the past medical information included in the right canine 508 is the same as the tooth in the inference result. Since both the past medical information and the inference result are bR3 and match, the process proceeds to step S714.

[0159] In step S714, the CPU 201 stores the tooth number and state of the inference result classified as the right canine 508 as corrected. A circle is temporarily stored in row 816 of bR3 in the correction column 804.

[0160] In step S715, if there is no contradiction between the tooth condition in the past medical record for the right canine 505 and the inference result or the result of determining that it is an artificial tooth, the CPU 201 stores the corrected information. Since bR3 is a bridge in the past medical record and a sound tooth in the inference result, there is no contradiction and the condition in the past medical record is stored. This is as shown in row 816 of bR3 in correction column 804.

[0161] In step S716, the CPU 201 increments the counter n for executing processing for each classification.

[0162] [Right premolar] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 6, it is determined that the counter n is equal to or less than max, and the process proceeds to step S705.

[0163] In step S705, CPU 201 determines whether the number of teeth classified into sixth tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in right premolar 509 in the past medical information is one, and the inference result is two, it is determined that there is a mismatch, and the process proceeds to step S709.

[0164] In step S709, CPU 201 determines whether the number of teeth classified into n-th tooth category 502 has increased in the inference result compared to the past medical information. Since the number of teeth included in right premolar 509 in the past medical information is one, but the inference result is two, it is determined that the number has increased, and the process proceeds to step S710.

[0165] In step S710, the CPU 201 checks the consistency of the past medical records with the inference results and then determines which teeth have been added. The tooth number in the past medical information included in the right premolar 509 is only bR5, whereas the tooth numbers in the inference results are bR4 and bR5. Furthermore, because bR4 is missing and a pontic in the past medical records, it is determined that bR4 is the tooth that has been added.

[0166] In step S711, the CPU 201 determines that teeth added without past treatment records may be artificial teeth such as implants. However, since bR4 is known to be an artificial pontic tooth from past medical records, it is not subject to notification.

[0167] In step S708, the CPU 201 stores the tooth number and state of the inference result classified as right premolar 509 as corrected. Since bR4 is known to be a pontic in step S711, it becomes as shown in row 817 of bR4 in correction column 804. Since bR5 is inferred to be a sound tooth, a circle is temporarily recorded in row 818 of bR5 in correction column 804.

[0168] In step S715, if there is no contradiction between the tooth condition of the right premolar 509 in the past medical records and the inference result or the result of determining that it is an artificial tooth, the CPU 201 stores it as corrected information. bR4 is a pontic in the past medical records and has been determined to be an artificial tooth in step S711, so there is no contradiction and the state of the past medical records is stored. bR5 is a bridge in the past medical records and the inference result is that it is a sound tooth, so there is no contradiction and the state of the past medical records is stored. The result is as shown in row 817 of bR4 and row 818 of bR5 in the correction column 804 and notification column 805.

[0169] In step S716, the CPU 201 increments a counter n for executing processing for each classification.

[0170] [Right molar] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 7, it is determined to be equal to or less than max, and the process proceeds to step S705.

[0171] In step S705, CPU 201 determines whether the number of teeth classified into seventh tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in right molar 510 in the past medical information is three and the inference result is two, it is determined that there is no match, and the process proceeds to step S709.

[0172] In step S709, CPU 201 determines whether the number of teeth classified as right molar 510 has increased in the inference result compared to the past medical information. The number of teeth included in right molar 510 in the past medical information is three, whereas the inference result shows two, so it is determined that the number has decreased and the process proceeds to step S712.

[0173] In step S712, CPU 201 compares the past medical results with the inference results, checks the consistency between the past medical records and the inference results, and then extracts the number of the removed tooth. The tooth numbers in the past medical information included in right molar 510 are bR6, bR7, and bR8. The tooth numbers in the inference results are also bR6 and bR8. If the extracted teeth are unknown, for convenience they will be bR6 and bR7. Furthermore, the tooth condition in row 819 of bR6 in past record column 802 is FMC, and the tooth condition in row 819 of bR6 in inference result column 803 is also FMC, which matches. Therefore, it is determined that the past medical records and the inference results for bR6 match. Next, the tooth condition of row 820 of bR7 in the past record column 802 is a partial metal restoration, row 820 of bR7 in the inference result column 803 is a missing tooth, and row 821 of bR8 is a partial metal restoration, so the inference result bR8 is corrected to bR7. Then, bR8 is determined to be missing.

[0174] In step S713, the CPU 201 determines that teeth that are missing without treatment records in past medical results may have been lost due to extraction, fracture, periodontal disease, etc. Therefore, they are associated with the tooth number and made the target of notification. Since bR8 is missing without any past medical records, row 821 of bR4 in the notification column 805 is marked as notified.

[0175] In step S708, the CPU 201 stores the tooth number and state of the inference result classified as the right molar 510 as corrected. bR6 is inferred to be FMC, and bR8 is changed to bR7, with the state being partially metal restored. Therefore, the results are as shown in row 819 of bR6, row 820 of bR7, and row 820 of bR8 in the correction column 804.

[0176] In step S715, the CPU 201 stores the tooth state in the past medical records classified as right molar 510 as having been corrected if it is consistent with the inference result or the result of determining that it is an artificial tooth. bR6 and bR7 have been corrected in step S712 so that their states in the past medical records and the inference result match. Therefore, there is no contradiction, and step S708 remains unchanged. bR8 is a sound tooth in the past medical records, has been determined to be missing in step S712, and is also the target of notification, so it remains unchanged in step S708.

[0177] In step S716, the CPU 201 increments a counter n for executing processing for each classification.

[0178] In step S704, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 8, it is determined that the counter n is greater than max, and the process proceeds to step S717.

[0179] In step S717, CPU 201 references the corrected tooth number and condition, and if the condition includes a disease, sets that tooth number as the notification target. In correction column 804, bL1, bL2, and bR1 are acid erosion, and bL7 is caries and a disease, so row 806 for bL1, row 807 for bL2, row 812 for bL7, and row 814 for bR1 in notification column 805 are set to include notification. This completes the process.

[0180] [Effect description] As described above, according to the image processing device of this embodiment, even if a patient has congenitally missing teeth and has assigned tooth numbers for convenience, when inferring tooth numbers and tooth conditions from an image, past medical information is used to correct the tooth numbers to the conveniently assigned tooth numbers for each tooth classification. This allows consistency with past medical information. After consistency is achieved, the tooth conditions that differ from the past medical information are output, resulting in a flawless and accurate output result. Furthermore, since notification is made after consistency is achieved, it is easier to grasp changes. Even in cases where it is difficult to distinguish between healthy teeth and artificial teeth from visible light images, if past medical information contains records of artificial teeth, etc., consistency can be achieved, allowing the presence or absence of artificial teeth or teeth to be correctly determined as a corrected result. In addition, if an unnatural tooth appears, the possibility of an artificial tooth can be notified. If an unnatural tooth is missing, the possibility of tooth extraction can be notified. If past medical examination results already indicate that the tooth is an artificial tooth and there are no notable changes, unnecessary notifications can be avoided. In this embodiment, notification is given when there is an increase or decrease in the number of teeth, but notification can also be given when there is an unnatural mismatch, such as when a treated tooth has been replaced with a normal tooth, making it possible to detect changes that require attention, such as when a partially restored metal tooth is replaced with a normal tooth. This can prevent patient mix-ups.

[0181] Furthermore, in this embodiment, an intraoral photograph of the mandibular occlusal view is used as an example, but even if inference is made using images of the maxillary occlusal view, front view, or left and right lateral view, by correcting using past medical examination results in the manner described above, it is possible to notify of changes that require attention while ensuring consistency with the past medical examination results.

[0182] (Second embodiment) In this embodiment, we will focus on intraoral images of the upper jaw of a child during the period when both baby teeth and permanent teeth are present, and explain a method for correcting the inference results so that they are consistent with past medical information based on specified rules described below.

[0183] 9 is a flowchart showing the process of determining a partially impacted tooth as an erupted tooth in this embodiment. In the inference using machine learning of this embodiment, the same process as in FIG. 6 will be described with reference to the corresponding steps in FIG. 6, and the differences will be mainly explained.

[0184] Step S901 is similar to step S601.

[0185] Step S902 is similar to step S602.

[0186] In step S903, the CPU 201 assigns 0 to a temporary number tmpTN for assigning a temporary management number to a tooth determined to be an erupted tooth.

[0187] Step S904 is similar to step S603.

[0188] In step S905, the CPU 201 determines whether the cnt-th state is a partially impacted tooth. The order in which the states are read is the same as the order described in step S604.

[0189] Step S906 is similar to step S604.

[0190] Step S907 is similar to step S605.

[0191] Step S908 is similar to step S606.

[0192] Step S909 is similar to step S607.

[0193] Step S910 is similar to step S608.

[0194] Step S911 is similar to step S609.

[0195] In step S912, the CPU 201 increments the temporary number tmpTN.

[0196] In step S913, the CPU 201 stores the coordinates of the tooth detected as being erupted in association with the tmpTN-th tooth.

[0197] 10A and 10B are flowcharts showing the correction process for mixed dentition in this embodiment.

[0198] 11 is an age-to-dental correspondence table showing age and the possibility of the presence or absence of teeth in this embodiment. In this embodiment, a table for upper jaw teeth is shown, but it is assumed that there is a similar table for lower jaw teeth.

[0199] A method for correcting the inference result using past medical information and ensuring consistency with the past medical information will be described using Figures 10A, 10B, and 11. The same processes as in Figure 7 are assigned the corresponding step numbers in Figure 7, and the differences will be mainly described.

[0200] In step S1000, the CPU 201 loads the age-tooth correspondence table into the RAM 203. The age-tooth correspondence table indicates the possibility of having teeth according to age. Details will be described later.

[0201] Step S1001 is the same as step S700. In the first embodiment, since the target is an intraoral image of the permanent dentition period, rows 504 to 510 in Fig. 5 are classified only into rows relating to permanent teeth, but when the target is an intraoral image of the mixed dentition period, rows 511 to 515 relating to baby teeth are also classified.

[0202] In step S1002, the CPU 201 provisionally determines whether the erupted tooth is a baby tooth or a permanent tooth. Potentially existing teeth are identified from the age-tooth correspondence table of FIG. 11 according to the patient's age. Then, teeth included in the inference result are excluded from the candidates, and a provisional determination is made based on the detected coordinates and the tooth coordinates in the inference result. If uL1 and uR1 have been inferred and the erupted tooth is between the coordinates of uL1 and uLC, it is provisionally determined to be uL2. If the erupted tooth is located at the coordinates behind uRE, it is provisionally determined to be uR6.

[0203] In FIG. 11, ◯ indicates that there is a high possibility that teeth are present, △ indicates that the possibility of having teeth is uncertain, and × indicates that there is a high possibility that teeth are absent.

[0204] 1101 is the 1-year-old column, which indicates the possibility of the presence or absence of teeth at age 1.

[0205] 1102 is the 2-year-old column, which indicates the possible presence or absence of teeth at age 2.

[0206] 1103 is the 3-year-old column, which indicates the possible presence or absence of teeth at age 3.

[0207] 1104 is the 4-year-old column, which indicates the possible presence or absence of teeth at age 4.

[0208] 1105 is the 5-year-old column, which indicates the possible presence or absence of teeth at age 5.

[0209] 1106 is the 6-year-old column, which indicates the possible presence or absence of teeth at age 6.

[0210] 1107 is the 7-year-old column, which indicates the possible presence or absence of teeth at age 7.

[0211] 1108 is the 8-year-old column, which indicates the possible presence or absence of teeth at age 8.

[0212] 1109 is the 9-year-old column, which indicates the possible presence or absence of teeth at age 9.

[0213] 1110 is the 10-year column, indicating the possible presence or absence of teeth at age 10.

[0214] 1111 is the 11-year-old column, indicating the possible presence or absence of teeth at age 11.

[0215] 1112 is the 12-year-old column, indicating the possible presence or absence of teeth at age 12.

[0216] 1113 is the 13-year-old column, indicating the possible presence or absence of teeth at age 13.

[0217] 1114 is the 14 to 16 years old column, which indicates the possible presence or absence of teeth at ages 14 to 16.

[0218] 1115 is the 17 to 21 years old column, which shows the possibility of the presence or absence of teeth between the ages of 17 and 21.

[0219] 1116 is the uLA line, which shows the possibility of presence or absence of uLA at each age.

[0220] 1117 is the uLB line, which indicates the possibility of presence or absence of uLB at each age.

[0221] 1118 is the uLC line, which indicates the possibility of presence or absence of uLC at each age.

[0222] 1119 is the uLD line, which shows the possibility of presence or absence of uLD at each age.

[0223] 1120 is the uLE line, which shows the possibility of presence or absence of uLE at each age.

[0224] 1121 is the uRA line, which shows the possibility of presence or absence of uRA at each age.

[0225] 1122 is the uRB line, which indicates the possibility of the presence or absence of uRB at each age.

[0226] 1123 is the uRC row, which indicates the possibility of presence or absence of uRC at each age.

[0227] 1124 is the uRD line, which indicates the possibility of presence or absence of uRD at each age.

[0228] 1125 is the uRE line, which indicates the possibility of presence or absence of uRE at each age.

[0229] 1126 is the uL1 line, which shows the possibility of presence or absence of uL1 at each age.

[0230] 1127 is the uL2 line, which shows the possibility of presence or absence of uL2 at each age.

[0231] 1128 is the uL3 line, which shows the possibility of presence or absence of uL3 at each age.

[0232] 1129 is the uL4 line, which shows the possibility of presence or absence of uL4 at each age.

[0233] 1130 is the uL5 line, which shows the possibility of presence or absence of uL5 at each age.

[0234] 1131 is the uL6 line, which shows the possibility of presence or absence of uL6 at each age.

[0235] 1132 is the uL7 line, which shows the possibility of presence or absence of uL7 at each age.

[0236] 1133 is the uL8 line, which shows the possibility of presence or absence of uL8 at each age.

[0237] 1134 is the uR1 line, which shows the possibility of presence or absence of uR1 at each age.

[0238] 1135 is the uR2 line, which shows the possibility of presence or absence of uR2 at each age.

[0239] 1136 is the uR3 line, which shows the possibility of presence or absence of uR3 at each age.

[0240] 1137 is the uR4 line, which shows the possibility of presence or absence of uR4 at each age.

[0241] 1138 is the uR5 line, which shows the possibility of presence or absence of uR5 at each age.

[0242] 1139 is the uR6 line, which shows the possibility of presence or absence of uR6 at each age.

[0243] 1140 is the uR7 line, which shows the possibility of presence or absence of uR7 at each age.

[0244] 1141 is the uR8 line, which shows the possibility of presence or absence of uR8 at each age.

[0245] In step S1003, the CPU 201 classifies the tooth numbers of the inference results stored in the RAM 203 and the tooth numbers provisionally determined as erupted teeth into the categories explained in FIG.

[0246] In step S1004, the CPU 201 substitutes the number of classifications that have one or more classified teeth based on the past medical information and the inference results into a value max used to determine the termination condition.

[0247] Step S1005 is similar to step S703.

[0248] In step S1006, CPU 201 determines whether counter n is equal to or less than max. If counter n is equal to or less than max, the process proceeds to step S1007. If counter n is greater than max, the correction process for each classification is considered to have ended, and the process proceeds to step S1018.

[0249] In step S1007, CPU 201 determines whether the number of teeth classified into n-th tooth category 502 matches between the past medical information and the inference result. If they match, the process proceeds to step S1008. If they do not match, the process proceeds to step S1010.

[0250] Step S1008 is the same as step S706. If they are the same tooth, the process proceeds to step S1017. If they are different teeth, the process proceeds to step S1009.

[0251] Step S1009 is the same as step S707. For the primary incisor 511, correction is made so that the central coordinates of the rectangles of the tooth numbers in the X-axis direction are in ascending order uRB, uRA, uLA, uLB and so that they match the past medical examination results.

[0252] Step S1010 is the same as step S709. If the number of teeth has increased, the process proceeds to step S1011. If the number of teeth has decreased, the process proceeds to step S1013.

[0253] Step S1011 is similar to step S710.

[0254] In step S1012, CPU 201 refers to the age-tooth correspondence table in RAM 203 and determines whether the tooth is likely to be increased. Referring to the patient's age column and the row of the corresponding tooth in FIG. 11, if it is marked with a circle or a triangle, it is determined that there is a possibility that the tooth will be increased. If it is marked with an ×, it is determined that there is no possibility that the tooth will be increased. If the tooth is likely to be increased, processing proceeds to step S1018. If the tooth is not likely to be increased, processing proceeds to step S1013.

[0255] In step S1013, the CPU 201 determines that the tooth number inference or the determination of the erupted tooth may be incorrect, and sets the tooth numbers determined to have increased in step S1011 as notification targets.

[0256] Step S1014 is similar to step S712.

[0257] In step S1015, CPU 201 refers to the age-tooth correspondence table in RAM 203 and determines whether the tooth is likely to be lost. Referring to the patient's age column and the row of the corresponding tooth in FIG. 11, if it is marked with a △ or ×, it is determined that there is a possibility that the tooth will be lost. If it is marked with a ◯, it is determined that there is no possibility that the tooth will be lost. If the tooth is likely to be lost, processing proceeds to step S1018. If the tooth is not likely to be lost, processing proceeds to step S1016.

[0258] In step S1016, the CPU 201 determines that the tooth number inference or the determination of the erupted tooth may be incorrect, and sets the tooth numbers determined to have decreased in step S1014 as notification targets.

[0259] Step S1017 is similar to step S714.

[0260] Step S1018 is similar to step S708.

[0261] Step S1019 is similar to step S715.

[0262] Step S1020 is the same as step S716, and the process then returns to step S1006.

[0263] In step S1021, CPU 201 determines whether or not there are any teeth that should not coexist. For this determination, a table for determining incompatible teeth, which will be described later, is used. If it is determined that there are any teeth that should not coexist, the process proceeds to step S1022. If it is determined that there are no teeth that should not coexist, the process proceeds to step S1023.

[0264] In step S1022, the CPU 201 sets the tooth numbers of the baby teeth that should not coexist as the notification targets.

[0265] Step S1023 is similar to step S717.

[0266] FIG. 12 is a determination table of teeth that should not coexist in this embodiment.

[0267] 1201 is the tooth number sequence, indicating the numbers of the maxillary and mandibular deciduous teeth.

[0268] 1202 is a sequence of numbers for incompatible teeth, indicating the numbers of the permanent teeth of the lower and upper jaws.

[0269] 1203 to 1212 are the number lines for each baby tooth, and indicate the correspondence between each baby tooth and the permanent teeth that should not coexist.

[0270] Figure 13 is a table showing an example of the correspondence between past medical information, inference results, and corrected results. Figure 8 shows the lower jaw teeth of the permanent dentition, while Figure 13 shows an example of the upper jaw teeth of the mixed dentition. Corresponding numbers are used for information that is the same as Figure 8, and the differences are mainly explained.

[0271] 1301 is the same as 801, 1302 is the same as 802, 1303 is the same as 803, 1304 is the same as 804, and 1305 is the same as 805.

[0272] Lines 1306 to 1331 store information about each tooth. Figure 8 shows an example of only the permanent dentition, so only the permanent teeth are included, whereas in the case of a mixed dentition, the primary teeth are included. Also, since this is the upper jaw, the tooth numbers begin with the letter u.

[0273] 1306 is a line that holds information about uLA, 1307 is a line that holds information about uLB, 1308 is a line that holds information about uLC, 1309 is a line that holds information about uLD, 1310 is a line that holds information about uLE, 1311 is a line that holds information about uRA, 1312 is a line that holds information about uRB, 1313 is a line that holds information about uRC, 1314 is a line that holds information about uRD, and 1315 is a line that holds information about uRE.

[0274] 1316 is a row that holds information about uL1, 1317 is a row that holds information about uL2, 1318 is a row that holds information about uL3, 1319 is a row that holds information about uL4, 1320 is a row that holds information about uL5, 1321 is a row that holds information about uL6, 1322 is a row that holds information about uL7, 1323 is a row that holds information about uL8, 1324 is a row that holds information about uR1, 1325 is a row that holds information about uR2, 1326 is a row that holds information about uR3, 1327 is a row that holds information about uR4, 1328 is a row that holds information about uR5, 1329 is a row that holds information about uR6, 1330 is a row that holds information about uR7, and 1331 is a row that holds information about uR8.

[0275] The results of the correction process will be described with reference to Figures 10A, 10B, 11, and 13. In this embodiment, the patient is assumed to be 8 years old.

[0276] In step S1000, the CPU 201 reads the age-to-tooth correspondence table into the RAM 203. Since the patient is 8 years old, the 8-year-old column 1108 in Fig. 11 is referenced and the possibility of the presence or absence of teeth at the age of 8 is used in the subsequent processing.

[0277] In step S1001, the CPU 201 classifies the tooth numbers present in the past medical information stored in the RAM 203 into the categories explained in Fig. 5. When the information in the past record sequence 1302 is divided, the primary incisors 511 are uLB and uRA, the left primary canine 512 is uLC, the left primary molar 513 is uLD and uLE, the right primary canine 514 is uRC, the right primary molar 515 is uRD and uRE, and there are no permanent teeth.

[0278] In step S1002, the CPU 201 provisionally determines whether the erupted tooth is a baby tooth or a permanent tooth. Potentially existing teeth are identified from an age-tooth correspondence table based on the patient's age. Referring to the 8-year-old column 1108, all baby teeth are marked with either a circle or a triangle, indicating that uLA, uLB, uRA, and uRB, which are marked with a triangle, are in the process of falling out. Furthermore, for permanent teeth, uL1, uL2, uR1, and uR2 are marked with a triangle, indicating that they are in the process of erupting. uL6 and uR6 are marked with a circle, indicating that they are likely to have already erupted. Referring to the inference result column 1303 based on this information, uL1 and uL2 have been detected, but uLA, uLB, and uRA have not. In other words, they are in the missing state. Assuming that the coordinates of the partially impacted tooth are near uRB and uL1, it can be provisionally determined to be uR1.

[0279] In step S1003, the CPU 201 divides the tooth numbers of the inference results stored in the RAM 203 and the tooth numbers provisionally determined as erupted teeth into the categories explained in Fig. 5. When the information in the inference result column 1303 is divided, the incisor 504 is uL1, uL2, and uR1, the left molar 507 is uL6, the right molar 510 is uR7, the deciduous incisor 511 is uRB, the left deciduous canine 512 is uLC, the left deciduous molar 513 is uLD and uLE, the right deciduous canine 514 is uRC, and the right deciduous molar 515 is uRD and uRE.

[0280] In step S1004, the CPU 201 assigns the number of categories that have one or more teeth classified based on the past medical information and the inference results to the value max used to determine the termination condition. Since there are now eight categories: deciduous incisor 511, left deciduous canine 512, left deciduous molar 513, right deciduous canine 514, right deciduous molar 515, incisor 504, left molar 507, and right molar 510, max is 8.

[0281] In step S1005, the CPU 201 assigns 1 to a counter n for executing processing for each classification.

[0282] [Incisor] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 1, it is determined that the counter n is equal to or less than max, and the process proceeds to step S1007.

[0283] In step S1007, CPU 201 determines whether the number of teeth classified into first tooth category 502 matches the past medical information and the inference result. Since the number of teeth included in incisor 504 in the past medical information is 0 and the inference result is 3, it is determined that there is no match, and the process proceeds to step S1010.

[0284] In step S1010, CPU 201 determines whether the number of teeth classified as incisors 504 has increased in the inference result compared to the past medical information. Since the number has increased from 0 to 3, it is determined that the number has increased, and the process proceeds to step S1011.

[0285] In step S1011, the CPU 201 determines which teeth have been added by verifying the consistency of the past medical records and the inference results. Incisor 504 does not exist in the past medical information, and the tooth numbers of the inference results are uL1, uL2, and uR1. It can also be seen from the age-tooth correspondence table that an 8-year-old patient is in the period when uL1, uL2, uR1, and uR2 will erupt. Therefore, the tooth numbers of the erupted teeth that have been inferred and provisionally determined are used, and uL1, uL2, and uR1 are determined to be the new teeth.

[0286] In step S1012, CPU 201 determines whether or not there is a possibility that the tooth will be added, by referring to the age-tooth correspondence table in RAM 203. In the age-tooth correspondence table, row 1126 of uL1, row 1127 of uL2, and row 1134 of uR1 in 8-year-old column 1108 are all △, so it is determined that there is a possibility that the tooth will be added, and the process proceeds to step S1018.

[0287] In step S1018, the CPU 201 stores the tooth number and state of the inference result classified as the incisor 504 as corrected, as shown in row 1316 of uL1, row 1317 of uL2, and row 1324 of uR1 in the correction column 1304.

[0288] In step S1019, the CPU 201 stores the state of the tooth in the past medical record classified as incisor 504 as corrected if it is consistent with the inference result or the result of determining that the tooth is an artificial tooth. Since the tooth was not yet grown in the past and is in the period of replacement, there is no contradiction with the inference result, and the state remains as it is in step S1018.

[0289] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 2, and returns the process to step S1006.

[0290] [Left molar] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 2, it is determined to be equal to or less than max, and the process proceeds to step S1007.

[0291] In step S1007, CPU 201 determines whether the number of teeth classified into second tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in left molar 507 in the past medical information is 0 and the inference result is 1, it is determined that there is no match, and the process proceeds to step S1010.

[0292] In step S1010, CPU 201 determines whether the number of teeth classified as left molars 507 has increased in the inference result compared to the past medical information. Because the number has increased from 0 to 1, it is determined that the number has increased, and the process proceeds to step S1011.

[0293] In step S1011, the CPU 201 determines which teeth have been added by verifying the consistency of the past medical records and the inference results. The left molar 507 does not exist in the past medical information, and the tooth number of the inference result is uL6. It can also be seen from the age-tooth correspondence table that an 8-year-old patient is in the period when uL6 should have erupted. Therefore, the inferred tooth number is used, and uL6 is determined to be the tooth that has been added.

[0294] In step S1012, CPU 201 determines whether or not there is a possibility that the tooth will be added, by referring to the age-tooth correspondence table in RAM 203. Since row 1131 of uL6 in column 1108 of age 8 in the age-tooth correspondence table is "○", it is determined that there is a possibility that the tooth will be added, and the process proceeds to step S1018.

[0295] In step S1018, the CPU 201 stores the tooth number and state of the inference result classified as the left molar 507 as corrected, as shown in row 1321 of uL6 in the correction column 1304.

[0296] In step S1019, the CPU 201 stores the tooth status in the past medical records classified as left molar 507 as corrected if it is consistent with the inference result or the result of determining that it is an artificial tooth. Since the tooth was not yet grown in the past and is in the period of replacement, there is no contradiction with the inference result, so the tooth is left as corrected and remains as is in step S1018.

[0297] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 3, and returns the process to step S1006.

[0298] [Right molar] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 3, it is determined that the counter n is equal to or less than max, and the process proceeds to step S1007.

[0299] In step S1007, CPU 201 determines whether the number of teeth classified into third tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in right molar 510 in the past medical information is 0 and the inference result is 1, it is determined that there is no match, and the process proceeds to step S1010.

[0300] In step S1010, CPU 201 determines whether the number of teeth classified as right molars 510 has increased in the inference result compared to the past medical information. Because the number has increased from 0 to 1, it is determined that the number has increased, and the process proceeds to step S1011.

[0301] In step S1011, the CPU 201 determines which teeth have been added by verifying the consistency of the past medical records and the inference results. The right molar 510 does not exist in the past medical information, and the tooth number of the inference result is uR7. However, it can be read from the age-tooth correspondence table that for an 8-year-old patient, uR7 is not in the period when it should erupt, but uR6 is in the period when it should erupt. Therefore, the inferred tooth number is corrected and uR6 is determined to be the tooth that has been added.

[0302] In step S1012, CPU 201 determines whether or not there is a possibility that the tooth will be added, by referring to the age-tooth correspondence table in RAM 203. Since row 1131 of uR6 in column 1108 of age 8 in the age-tooth correspondence table is "○", it is determined that there is a possibility that the tooth will be added, and the process proceeds to step S1018.

[0303] In step S1018, the CPU 201 stores the tooth number and state of the inference result classified as the right molar 510 as corrected. This is as shown in row 1329 of uR6 in the correction column 1304.

[0304] In step S1019, the CPU 201 stores the tooth state in the past medical records classified as right molar 510 as corrected if it is consistent with the inference result or the result of determining that it is an artificial tooth. Since the tooth was not yet grown in the past and is in the period of replacement, there is no contradiction with the inference result, so the tooth remains corrected and remains as it is in step S1018.

[0305] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 4, and returns the process to step S1006.

[0306] [Baby incisors] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 4, it is determined to be equal to or less than max, and the process proceeds to step S1007.

[0307] In step S1007, CPU 201 determines whether the number of teeth classified into fourth tooth category 502 matches the past medical information and the inference result. Since the number of teeth included in primary incisor 511 in the past medical information is two, and the inference result is one, it is determined that there is no match, and the process proceeds to step S1010.

[0308] In step S1010, CPU 201 determines whether the number of teeth classified as primary incisors 511 has increased in the inference result compared to the past medical information. Because the number has decreased from two to one, it is determined that the number has decreased, and the process proceeds to step S1014.

[0309] In step S1014, the CPU 201 determines which teeth have been lost after verifying the consistency of the past medical records with the inference results. The tooth numbers in the past medical information included in the primary incisor 511 are uLB and uRA. The tooth number in the inference results is uRB. uRB does not exist in the past medical records. If either uRA or uRB is a congenitally missing tooth, then for convenience, it will be uRA. It can also be seen from the age-tooth correspondence table that an 8-year-old patient is at the age where uLB will be lost. Therefore, it is determined that uLB is missing.

[0310] In step S1015, CPU 201 determines whether the tooth is likely to be lost by referring to the age-tooth correspondence table in RAM 203. Since row 1117 of uLB in 8-year-old column 1108 in the age-tooth correspondence table is △, it is determined that there is a possibility that the tooth will be lost, and the process proceeds to step S1018.

[0311] In step S1018, the CPU 201 stores the tooth number and state of the inference result classified as primary incisor 511 as having been corrected. Since the tooth inferred as uRB in step S1014 has been changed to uRA, the state of uRB is associated with uRA, and uRA becomes a sound tooth. This is as shown in row 1311 of uRA in correction column 1304. Furthermore, since uLA, uLB, and uRB have been determined to be missing, this is as shown in row 1306 of uLA, row 1307 of uLB, and row 1308 of uRB in correction column 1304.

[0312] In step S1019, the CPU 201 stores the state of the tooth in the past medical record classified as the primary incisor 511 as corrected if there is no contradiction with the inference result or the result of determining that the tooth is an artificial tooth. Since uRA is a sound tooth in both the past medical record and the inference result, no contradiction occurs, and the state remains as corrected in step S1018.

[0313] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 5, and returns the process to step S1006.

[0314] [Left deciduous canine] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 5, it is determined that the counter n is equal to or less than max, and the process proceeds to step S1007.

[0315] In step S1007, CPU 201 determines whether the number of teeth classified into fifth tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in left deciduous canine 512 in the past medical information is one and the inference result is also one, it is determined that they match, and the process proceeds to step S1008.

[0316] In step S1008, the CPU 201 determines whether the tooth in the past medical information included in the left deciduous canine 512 is the same as the tooth in the inference result. Since both the past medical information and the inference result are uLC and match, the process proceeds to step S1017.

[0317] In step S1017, the CPU 201 stores the tooth number and state of the inference result classified as the left deciduous canine 512 as corrected. This is as shown in row 1308 of uLC in the correction column 1304.

[0318] In step S1019, the CPU 201 stores as corrected the tooth states in the past medical records classified as the left deciduous canine 512 that are consistent with the state in the inference result or have been determined to be subject to notification. Since the uLC is a healthy tooth in both the past medical records and the inference result, and no contradiction occurs, it remains as corrected in step S1017.

[0319] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 6, and returns the process to step S1006.

[0320] [Left baby molar] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 6, it is determined that the counter n is equal to or less than max, and the process proceeds to step S1007.

[0321] In step S1007, CPU 201 determines whether the number of teeth classified into sixth tooth category 502 matches the past medical information and the inference result. Since the number of teeth included in left deciduous molar 513 in the past medical information is two and the inference result is also two, it is determined that they match and the process proceeds to step S1008.

[0322] In step S1008, the CPU 201 determines whether the tooth in the past medical information included in the left deciduous molar 513 is the same as the tooth in the inference result. Since both the past medical information and the inference result are uLD and uLE and match, the process proceeds to step S1017.

[0323] In step S1017, the CPU 201 stores the tooth number and state of the inference result classified as the left primary molar 513 as corrected. The state of uLD is stored as a sound tooth, and the state of uLE is stored as a decayed tooth. This is as shown in row 1309 of uLD and row 1310 of uLE in the correction column 1304.

[0324] In step S1019, the CPU 201 stores as corrected the tooth states in the past medical records classified as left primary molar 513 that are consistent with the state of the inference result or have been determined to be subject to notification. In the past medical records, both uLD and uLE are sound teeth, and in the inference result, uLD is sound and uLE is decayed. Because sound teeth can develop cavities, there is no contradiction, and step S1018 remains unchanged.

[0325] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 7, and returns the process to step S1006.

[0326] [Right deciduous canine] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 7, it is determined to be equal to or less than max, and the process proceeds to step S1007.

[0327] In step S1007, CPU 201 determines whether the number of teeth classified into seventh tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in right deciduous canine 514 in the past medical information is one and the inference result is also one, it is determined that they match, and the process proceeds to step S1008.

[0328] In step S1008, the CPU 201 determines whether the tooth in the past medical information included in the right deciduous canine 514 is the same as the tooth in the inference result. Since both the past medical information and the inference result are uRCs and match, the process proceeds to step S1017.

[0329] In step S1017, the CPU 201 stores the tooth number and state of the inference result classified as the right deciduous canine 514 as corrected. This is as shown in row 1313 of uRC in the correction column 1304.

[0330] In step S1019, the CPU 201 stores as corrected the tooth states in the past medical records classified as the right deciduous canine 514 that are consistent with the state in the inference result or have been determined to be subject to notification. Since the tooth is healthy in both the past medical records and the inference result, the uRC determines that there is no contradiction and leaves step S1017.

[0331] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 8, and returns the process to step S1006.

[0332] [Right deciduous molar] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 8, it is determined to be equal to or less than max, and the process proceeds to step S1007.

[0333] In step S1007, CPU 201 determines whether the number of teeth classified into eighth tooth category 502 matches between the past medical information and the inference result. Since the number of teeth included in right deciduous molar 515 in the past medical information is two and the inference result is also two, it is determined that they match and the process proceeds to step S1008.

[0334] In step S1008, the CPU 201 determines whether the tooth in the past medical information included in the right deciduous molar 515 is the same as the tooth in the inference result. Since both the past medical information and the inference result are uRD and uRE and match, the process proceeds to step S1017.

[0335] In step S1017, the CPU 201 stores the tooth number and state of the inference result classified as the right primary molar 515 as corrected. The state of uRD is stored as partial crown color restoration, and the state of uRE is stored as sound tooth. This is as shown in row 1314 of uRD and row 1315 of uRE in the correction column 1304.

[0336] In step S1019, the CPU 201 stores as corrected the tooth states in the past medical records classified as the right primary molar 515 that are consistent with the state of the inference result or have been determined to be subject to notification. In the past medical records, both uRD and uRE were healthy teeth, but in the inference result, uRD is partially restored and uRE is a healthy tooth. Since a healthy tooth can be partially restored, there is no contradiction, and step S1017 remains unchanged.

[0337] In step S1020, the CPU 201 increments the counter n, which is used to execute processing for each classification, to 9, and returns the process to step S1006.

[0338] In step S1006, the CPU 201 determines whether the counter n is equal to or less than max. Since the counter n is 9, it is determined that the counter n is greater than max, and the process proceeds to step S1021.

[0339] In step S1021, CPU 201 determines whether or not there are any teeth that should not coexist. In correction column 1304, the permanent teeth that are present are uL1, uL2, uL6, uR1, and uR6. By referring to tooth number column 1201 and incompatible tooth number column 1202 in the tooth determination table, it can be seen that uL1 should not coexist with uLA, uL2 should not coexist with uLB, and uR1 should not coexist with uRA, respectively. On the other hand, according to row 1311 for uRA in correction column 1304, both uRA and uR1 are present. Therefore, it is determined that there are any teeth that should not coexist, and processing proceeds to step S1022.

[0340] In step S1022, CPU 201 sets the tooth numbers of the baby teeth that should not coexist as notification targets, as shown in row 1311 of uRA in notification column 1305.

[0341] In step S1023, CPU 201 references the corrected tooth number and condition, and if the condition includes a disease, sets that tooth number as the notification target. In correction column 1304, uLE is a cavity and a disease, so row 1310 of uLE in notification column 1305 is set to "yes." This completes the process.

[0342] [Effect description] As described above, with the image processing device of this embodiment, even when a mixture of baby teeth and permanent teeth is present, when inferring tooth type and tooth condition from an image, past medical information is used to correct the tooth numbers assigned for convenience, and consistency with the past medical information is ensured. After consistency is ensured, the tooth condition that differs from the past medical information is output and notified, making it easier to grasp changes. Furthermore, by not notifying natural tooth replacements, taking into account factors such as the timing of tooth replacement, it becomes easier to grasp diseases that require more attention. Furthermore, it is possible to notify if permanent teeth grow in while baby teeth that should be removed remain, which could cause problems.

[0343] In this embodiment, an intraoral photograph of the maxillary occlusal view in the mixed dentition stage has been used as an example, but even when inference is made using images of the mandibular occlusal view, front view, or left and right lateral view, corrections can be made using past medical examination results in the above-described manner. This makes it possible to notify changes that require attention after ensuring consistency with past medical examination results.

[0344] (Third embodiment) In the first and second embodiments, examples were shown in which the past medical examination results only contained tooth numbers, the presence or absence of teeth, and treatment information, but if intraoral photographs and the coordinates of each tooth are recorded in the images, they may be referenced as past medical information. The relative positional relationship and presence or absence information of each tooth can be used to correlate with the results of inference using new intraoral images.

[0345] In this embodiment, a method for correcting the inference results based on the relative positional relationship of each tooth in the previous and new intraoral images and information on the presence or absence of teeth will be described.

[0346] 14 is a conceptual diagram showing the results of past medical examinations in this embodiment. It is assumed that the images are intraoral images from past medical examinations corresponding to the conceptual diagram of the inference results shown in FIG.

[0347] 1400 is an intraoral image taken during a previous medical examination. It is an intraoral image of the mandibular occlusion of an adult with only permanent teeth.

[0348] Reference numerals 1401 to 1412 denote rectangles associated with each tooth in past medical treatment, and each rectangle is associated with a label of the tooth number corresponding to the rectangle. Each rectangle is also associated with coordinate information on the image. The image processing device 101 requests and acquires past medical treatment information, including intraoral images, associated with the patient and stored in the medical treatment information DB 105 from the dental electronic medical record system 104 via the network 102, and the CPU 201 of the image processing device 101 stores the information in the HDD 205.

[0349] 1401 is a rectangle corresponding to the teeth of bR8.

[0350] 1402 is a rectangle corresponding to the teeth of bR7.

[0351] 1403 is a rectangle corresponding to the teeth of bR6.

[0352] 1404 is a rectangle corresponding to the teeth of bR5.

[0353] 1405 is a rectangle corresponding to the teeth of bR4.

[0354] 1406 is a rectangle corresponding to the teeth of bR3.

[0355] 1407 is a rectangle corresponding to the teeth of bR1.

[0356] 1408 is a rectangle corresponding to the teeth of bL1.

[0357] 1409 is a rectangle corresponding to the teeth of bL2.

[0358] 1410 is a rectangle corresponding to the teeth of bL3.

[0359] 1411 is a rectangle corresponding to the tooth bL5.

[0360] 1412 is a rectangle corresponding to the tooth bL7.

[0361] Figure 15 is a table showing an example of correspondence between past medical information, inference results, and corrected results. The past medical information is associated with Figure 14, and the inference results are associated with Figure 4. For information that is the same as in Figure 8, corresponding numbers are written, and the explanation will focus on the differences. 1501 is the same as 801.

[0362] Reference numeral 1502 denotes a past record column that holds past medical records. In 802, the presence or absence of teeth is indicated by a circle or an x, but in this embodiment, information on the corresponding rectangle in FIG. 14 is held. The state is the same as in 802.

[0363] 1503 is an inference result sequence that holds the tooth number and tooth status of the inference result. In 803, the presence or absence of a tooth is indicated by a circle or an x, but in this embodiment, information on the corresponding rectangle in Figure 4 is held. The status is the same as in 803.

[0364] Reference numeral 1504 denotes a post-correction column that holds the presence or absence and state of teeth after correction. In 804, the presence or absence of teeth is indicated by a circle or cross, but in this embodiment, information on the corresponding rectangle in FIG. 4 after correction is held. The state is the same as in 804.

[0365] 1505 is the same as 805.

[0366] Lines 1506 to 1521 are lines that hold information about each tooth. They are similar to lines 806 to 821.

[0367] A method for correcting the inference result using past medical information will be described with reference to FIGS. 4, 14, and 15.

[0368] The dental arch in the intraoral image from the previous medical record and the dental arch in the new intraoral image are resized and rotated so that they overlap by 80% or more, and if the rectangles of each tooth overlap by 80% or more, they are considered to be the same tooth. The dental arch is created by connecting the centers of the tooth rectangles with a Bézier curve. This line is made thicker by the average height and width of the tooth rectangles.

[0369] When the intraoral images are overlaid using the above method, it is clear that the tooth corresponding to rectangle 1401 of bR8 in the past medical results is missing. If rectangle 1402 of bR7 in the past medical results overlaps with rectangle 401 of bR8 in the inference results, rectangle 401 of the inference results is corrected to bR7. Using a similar method, rectangle 406 of the inference results is corrected to bR1, and rectangle 410 of the inference results is corrected to bL5. Since rectangle 411 of the inference results appears despite not being present in the past medical results, it is clear that it may be an artificial tooth. The inference status and the status in the past medical information are then associated, resulting in the result shown in correction column 1504. After that, notifications are made of diseases, unnaturally appearing teeth, and missing teeth. The result is then as shown in notification column 1505.

[0370] In this embodiment, the coordinates of the previous and new intraoral images were aligned by resizing and rotating the images so that the curves of the dental arches overlapped. However, if both the old and new intraoral images can be captured with the same angle of view and the same size using an imaging guide or the like, alignment is not necessary. Furthermore, the midline, the position of the incisors, the position of the molars, etc. may be resized and rotated to ensure that the rectangles of the teeth in the new and old images correspond to each other.

[0371] [Effect description] As described above, according to the image processing device of this embodiment, when inferring tooth type and tooth condition from an image, tooth numbers can be corrected based on the detected position of each tooth in a past image, and consistency with past medical information can be achieved. Then, after consistency is achieved, the tooth condition that is the difference from the past medical information can be notified, making it easier to grasp changes.

[0372] (Other embodiments) The present invention can also be realized by supplying a program that realizes one or more of the functions of the above-described embodiments to a system or device via a network or a storage medium, and having one or more processors in the computer of the system or device read and execute the program.The present invention can also be realized by a circuit (e.g., ASIC) that realizes one or more of the functions.

[0373] The disclosure of this specification includes the following image processing device, method, program, and storage medium.

[0374] (Item 1) a first acquisition means for acquiring past medical information of a target patient; a second acquisition means for acquiring an intraoral image of the patient; an inference means for inferring the type of tooth from the intraoral image; a correcting means for correcting the result of the inference by the inference means based on a predetermined rule so that the result is consistent with the past medical information; an output means for outputting information on the tooth type corrected by the correction means; An image processing device comprising:

[0375] (Item 2) 2. The image processing device according to item 1, wherein the tooth type information is the type of tooth and whether or not it is present.

[0376] (Item 3) 3. The image processing device according to item 1 or 2, wherein the past medical information is information on the types of teeth in past medical treatments and information on treatments for each tooth.

[0377] (Item 4) 4. The image processing device according to item 3, wherein the correction means corrects the result of the inference based on the information on the treatment.

[0378] (Item 5) 5. The image processing device according to any one of items 1 to 4, wherein the correction means determines the state of tooth loss associated with each type of tooth and corrects the type of tooth.

[0379] (Item 6) The image processing device described in any one of items 1 to 5, characterized in that the correction means corrects the tooth type so that the relative positional relationship of each tooth in the patient's past intraoral image is maintained.

[0380] (Item 7) 7. The image processing device according to any one of items 1 to 6, wherein the correction means has a third acquisition means for acquiring the condition of the patient, and corrects the result of the inference based on the condition of the patient.

[0381] (Item 8) 8. The image processing device according to item 7, wherein the condition of the patient is the age of the patient.

[0382] (Item 9) 9. The image processing device according to item 8, wherein the correction means corrects the result of the inference based on a correspondence table showing the age of the patient and the possibility of the presence or absence of teeth.

[0383] (Item 10) The image processing device described in any one of items 1 to 9 is characterized in that, when correcting the result of the inference, if there is an increase or decrease in teeth that violates the predetermined rules, the correction means notifies the user of this.

[0384] (Item 11) Item 11. The image processing device according to item 10, wherein the correction means notifies the user that there is a possibility of an artificial tooth if there is an increase in teeth that violates the predetermined rules.

[0385] (Item 12) Item 11. The image processing device according to item 10, wherein the correction means notifies the user that there is a possibility of tooth extraction if there is tooth loss that violates the predetermined rules.

[0386] (Item 13) Item 1. The image processing device according to item 1, further comprising a detection means for detecting the state of the teeth from the image, and a second output means for outputting the coordinates after correcting the tooth type and the detection result of the tooth state in association with each other.

[0387] (Item 14) a first acquisition step of acquiring past medical information of a target patient; a second acquisition step of acquiring an intraoral image of the patient; an inference step of inferring tooth types from the intraoral images; a correction step of correcting the result of the inference in the inference step based on a predetermined rule so that the result is consistent with the past medical information; an output step of outputting information on the tooth type corrected in the correction step; An image processing method comprising:

[0388] (Item 15) Item 15. A program for causing a computer to execute each step of the image processing method according to Item 14.

[0389] (Item 16) Item 15. A computer-readable storage medium storing a program for causing a computer to execute each step of the image processing method according to Item 14.

[0390] The invention is not limited to the above-described embodiments, and various changes and modifications can be made without departing from the spirit and scope of the invention. Accordingly, the following claims are appended to apprise the public of the scope of the invention. [Explanation of symbols]

[0391] 100: Dental system, 101: Image processing device, 102: Network, 103: Dental electronic medical record display terminal, 104: Dental electronic medical record system, 105: Medical information DB, 106: In-hospital system, 107: Patient information DB, 108: Imaging device, 109: Oral cavity

Claims

1. a first acquiring means for acquiring past medical information of a target patient; a second acquisition means for acquiring an intraoral image of the patient; an inference means for inferring the type of tooth from the intraoral image; a correcting means for correcting the result of the inference by the inference means based on a predetermined rule so that the result is consistent with the past medical information; an output means for outputting information on the tooth type corrected by the correction means; An image processing device comprising:

2. 2. The image processing device according to claim 1, wherein the tooth type information is the type of tooth and whether or not it is present.

3. 2. The image processing apparatus according to claim 1, wherein the past medical information includes information on the types of teeth and treatments for each tooth in past medical treatments.

4. 4. The image processing apparatus according to claim 3, wherein the correction means corrects the result of the inference based on the information on the treatment.

5. 2. The image processing device according to claim 1, wherein the correction means determines the state of tooth loss associated with each type of tooth and corrects the type of tooth.

6. The image processing device according to claim 1 , wherein the correction means corrects the tooth types so that the relative positional relationship of each tooth in the past intraoral image of the patient is maintained.

7. 2. The image processing apparatus according to claim 1, wherein the correction means comprises a third acquisition means for acquiring the state of the patient, and corrects the result of the inference based on the state of the patient.

8. 8. The image processing apparatus according to claim 7, wherein the condition of the patient is the age of the patient.

9. 9. The image processing apparatus according to claim 8, wherein said correcting means corrects the result of said inference based on a correspondence table showing the age of said patient and the possibility of the presence or absence of teeth.

10. 2. The image processing device according to claim 1, wherein the correction means notifies a user if there is an increase or decrease in teeth that violates the predetermined rules when correcting the result of the inference.

11. 11. The image processing device according to claim 10, wherein the correction means notifies the user that there is a possibility that the teeth are artificial teeth when there is an increase in the number of teeth that violates the predetermined rule.

12. 11. The image processing device according to claim 10, wherein the correction means notifies the user of the possibility of tooth extraction when there is a tooth loss that violates the predetermined rule.

13. 2. The image processing device according to claim 1, further comprising: a detecting means for detecting the state of the teeth from the image; and a second output means for outputting the coordinates after correcting the tooth type and the detection result of the tooth state in association with each other.

14. a first acquisition step of acquiring past medical information of a target patient; a second acquisition step of acquiring an intraoral image of the patient; an inference step of inferring tooth types from the intraoral images; a correction step of correcting the result of the inference in the inference step based on a predetermined rule so that the result is consistent with the past medical information; an output step of outputting information on the tooth type corrected in the correction step; An image processing method comprising:

15. A program for causing a computer to execute each step of the image processing method according to claim 14.

16. A computer-readable storage medium storing a program for causing a computer to execute each step of the image processing method according to claim 14.

Images