Orthodontic diagnosis and treatment report generation method and system based on artificial intelligence

Abstract

The invention relates to an artificial intelligence-based orthodontic diagnosis and treatment report generation method and system, and the method comprises the following steps: obtaining historical orthodontic related clinical data and corresponding diagnosis and treatment data, and carrying out the preprocessing; constructing a multi-task learning model based on artificial intelligence, inputting multi-modal data including clinical data and diagnosis and treatment data, and outputting a plurality of clinical diagnosis and treatment indications; constructing a causal model, inputting the clinical diagnosis and treatment indications and industry diagnosis and treatment standards, and outputting a treatment scheme corresponding to the clinical diagnosis and treatment indications; training a multi-task learning model and a causal learning model based on the acquired clinical data and diagnosis and treatment data; and obtaining orthodontic related clinical data of the current patient, preprocessing the orthodontic related clinical data, inputting the trained multi-task learning model and causal model to obtain an auxiliary diagnosis and treatment scheme, and generating a diagnosis and treatment report of the current patient. Compared with the prior art, the method has the advantages of high accuracy of generated auxiliary diagnosis and treatment results and the like.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based method and system for generating an oral orthodontic diagnosis and treatment report. Background Art

[0002] According to the oral disease survey report, nearly 90% of the world's population has some degree of oral problems, and many of them need dental treatment. At present, for oral diseases such as malocclusion, more and more patients choose to undergo oral orthodontics. However, the malocclusion of patients' teeth varies, and it is necessary to comprehensively consider the patient's dental condition, facial condition, CT imaging data, MRI data, etc. to give a comprehensive diagnosis report, which takes a lot of time for doctors. Relying solely on the doctor's judgment, different doctors are likely to give different diagnoses and treatment plans, and it is difficult to form a standard diagnosis report and treatment plan. Therefore, how to develop a system that can comprehensively consider the patient's clinical data and give accurate auxiliary diagnosis reports and auxiliary treatment plans is a difficult problem that needs to be faced in the clinic. With the development of artificial intelligence technology, it is possible to develop a system that can automatically generate oral orthodontic diagnosis reports and treatment recommendations based on existing clinical oral diagnosis and treatment big data. However, whether artificial intelligence can be used to obtain accurate and effective tooth-related features has a great impact on the accuracy of the generated diagnosis reports and treatment recommendations. The tooth-related features obtained by the current scheme are single and the coverage is not wide enough, resulting in low reliability of the generated results. Summary of the invention

[0003] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide an artificial intelligence-based orthodontic diagnosis and treatment report generation method and system. A multi-task deep learning network model is trained based on clinical multimodal data and expert knowledge to realize the automatic calculation of multiple parameters required in the orthodontic diagnosis and treatment process, and generate corresponding diagnosis reports and subsequent diagnosis and treatment recommendations, which can solve the problems of laborious and time-consuming diagnosis, inconsistent diagnostic standards and poor diagnostic accuracy in the prior art.

[0004] The purpose of the present invention can be achieved by the following technical solutions:

[0005] According to a first aspect of the present invention, a method for generating an orthodontic treatment report based on artificial intelligence is provided, the method comprising the following steps:

[0006] Acquiring historical orthodontic related clinical data and preprocessing the clinical data;

[0007] Acquiring diagnosis and treatment data corresponding to historical orthodontic-related clinical data, and preprocessing the diagnosis and treatment data;

[0008] Constructing a multi-task learning model based on artificial intelligence, wherein the input of the multi-task learning model is multimodal data, including pre-processed clinical data and diagnosis and treatment data, and the output of the multi-task learning model is multiple clinical diagnosis and treatment indications, including one or more of dental crowding, deep overbite, deep overjet, open bite, malocclusion type, skeletal and profile facial type;

[0009] Constructing a causal model, wherein the input of the causal model is the clinical diagnosis and treatment indication and the industry diagnosis and treatment standard, and the output of the causal model is the treatment plan corresponding to the clinical diagnosis and treatment indication;

[0010] Training the multi-task learning model and the causal learning model based on the acquired clinical data and diagnosis and treatment data;

[0011] The orthodontic-related clinical data of the current patient is obtained, and after preprocessing the clinical data, the trained multi-task learning model and causal model are input to obtain an auxiliary diagnosis and treatment plan, and a diagnosis and treatment report for the current patient is generated.

[0012] As a preferred technical solution, the orthodontic-related clinical data include at least two of the following: intraoral dental images, facial photos, oral scan data, dental CBCT data, head positioning frontal and lateral radiographs, oral panoramic radiographs, temporomandibular magnetic resonance imaging data, and medical history data.

[0013] As a preferred technical solution, the intraoral tooth images include at least two of the upper dentition images, lower dentition images, intraoral right side images, intraoral left side images, intraoral frontal images, facial front images, facial side images, and facial smiling images taken by a digital camera.

[0014] As a preferred technical solution, the multi-task learning model includes a multimodal feature extraction module and a multimodal feature fusion module. The multimodal feature extraction module uses a neural network to extract primary features and uses a large model to extract deep features. The multimodal feature fusion module fuses the features extracted by the multimodal feature extraction module according to the importance of the data modality.

[0015] As a preferred technical solution, the importance of the data modality is set through weights based on clinical diagnosis and treatment experience.

[0016] As a preferred technical solution, the preprocessing includes: denoising and standardizing facial photos, oral dental pictures or other image pictures; and extracting key text information from medical history materials and diagnosis and treatment data using natural language processing.

[0017] As a preferred technical solution, the diagnosis and treatment report includes patient information and auxiliary diagnosis and treatment recommendations, as well as at least one of the patient's facial photo analysis results, intraoral dental photo analysis results, dental image analysis results, and cephalometric analysis results, determined based on the input orthodontic-related clinical data type of the current patient.

[0018] As a preferred technical solution, the diagnosis and treatment report includes manually added feedback from professional dentists, and the feedback is used to dynamically optimize the multi-task learning model and causal model.

[0019] As a preferred technical solution, the crowding degree of the teeth represents the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch, which is divided into three degrees. Degree I crowding means that the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch is 2 to 4 mm, degree II crowding means that the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch is 4 to 8 mm; degree III crowding means that the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch is more than 8 mm, wherein the crown width is the distance between the most protruding points of the adjacent crown surfaces in the mesiodistal direction of the teeth, and the length of the existing arc of the dental arch is the distance from the mesial contact point of the first molar along the premolar surface to the cusp of the canine and then along the incisal edge of the maxillary incisor to the mesial contact point of the first molar on the opposite side.

[0020] According to a second aspect of the present invention, there is provided an artificial intelligence-based orthodontic treatment report generation system for implementing the above method, the system comprising:

[0021] Data acquisition module: acquiring historical orthodontic related clinical data and diagnosis and treatment data corresponding to the historical orthodontic related clinical data;

[0022] Data preprocessing module: preprocessing the clinical data and diagnosis and treatment data;

[0023] Model construction module: including a multi-task learning model construction submodule and a causal model construction submodule, wherein the multi-task learning model construction submodule is used to construct a multi-task learning model based on artificial intelligence, the input of the multi-task learning model is multimodal data, including pre-processed clinical data and diagnosis and treatment data, and the output of the multi-task learning model is a plurality of clinical diagnosis and treatment indications, including one or more of tooth crowding, deep overbite, deep overjet, open bite, malocclusion type, skeletal and profile facial types; the causal model construction submodule is used to construct a causal model, the input of the causal model is the clinical diagnosis and treatment indications and industry diagnosis and treatment standards, and the output of the causal model is the treatment plan corresponding to the clinical diagnosis and treatment indications;

[0024] Model optimization module: training the multi-task learning model and causal learning model based on the acquired clinical data and diagnosis and treatment data;

[0025] Report generation module: includes a diagnosis report generation submodule and a treatment report generation submodule, which are used to obtain the orthodontic-related clinical data of the current patient, and after pre-processing the clinical data, input the trained multi-task learning model and causal model. The diagnosis report generation submodule calls the output of the multi-task learning model to generate an auxiliary diagnosis report, and the treatment report generation submodule calls the output of the causal model to generate an auxiliary treatment plan, and generates a diagnosis and treatment report for the current patient based on the auxiliary diagnosis report and auxiliary treatment plan.

[0026] Compared with the existing technology, the present invention can make full use of multimodal data to identify clinical diagnosis and treatment indications, which can comprehensively and accurately characterize the relevant characteristics of oral orthodontics, thereby automatically generating accurate and effective diagnosis and treatment reports, helping dentists to quickly give diagnostic opinions, saving patients' outpatient time and improving the quality of medical treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flow chart of the method for generating an oral orthodontic treatment report of the present invention;

[0028] Figure 2 A schematic diagram of a diagnosis and treatment report generated in one embodiment of the present invention;

[0029] Figure 3 A structural diagram of the oral orthodontic diagnosis and treatment report generation system of the present invention;

[0030] Figure 4 A schematic diagram of the structure of an oral orthodontic treatment report generating terminal of the present invention. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0032] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0033] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantitative limitation, and may represent the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0034] Example 1

[0035] This embodiment provides a method for generating an oral orthodontic diagnosis and treatment report based on artificial intelligence, such as Figure 1 As shown, the method comprises the following steps:

[0036] Step 101, obtaining historical orthodontic related clinical data and preprocessing the clinical data.

[0037] In this embodiment, the orthodontic-related clinical data includes at least two of intraoral dental images, facial photos, oral scan data, dental CBCT data, head positioning frontal and lateral radiographs, oral panoramic radiographs, temporomandibular magnetic resonance imaging data, and medical history data. The intraoral dental images include at least two of the upper dentition images, lower dentition images, intraoral right side images, intraoral left side images, intraoral frontal images, facial frontal images, facial side images, and facial smile images taken by a digital camera.

[0038] Preprocessing includes: denoising and standardizing facial photos, oral dental images or other image images; using natural language processing (NLP) to extract key text information from medical history materials and diagnosis and treatment data.

[0039] Step 102: Acquire diagnosis and treatment data corresponding to historical orthodontic-related clinical data, and pre-process the diagnosis and treatment data.

[0040] Step 103, constructing a multi-task learning model based on artificial intelligence, the input of the multi-task learning model is multimodal data, including pre-processed clinical data and diagnosis and treatment data, and the output of the multi-task learning model is multiple clinical diagnosis and treatment indicators, including one or more of dental crowding, deep overbite, deep overjet, open bite, malocclusion type, skeletal and profile facial types.

[0041] The multi-task learning model includes a multimodal feature extraction module and a multimodal feature fusion module. The multimodal feature extraction module uses a neural network to extract primary features and uses a large model (such as ChatGPT, kimi, etc.) to extract deep features. The multimodal feature fusion module fuses the features extracted by the multimodal feature extraction module according to the importance of the data modality.

[0042] In this embodiment, the importance of data modalities is set by weights based on clinical diagnosis and treatment experience. The weights can be set to intraoral dental images (10%), facial photos (10%), oral scan data (10%), dental CBCT data (10%), head positioning frontal and lateral radiographs (15%), oral panoramic radiographs (15%), temporomandibular magnetic resonance imaging data (20%), and medical history data (10%).

[0043] The degree of crowding of teeth refers to the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch, which is generally divided into three degrees. Among them, for degree I crowding, the difference is 2 to 4 mm; for degree II crowding, the difference is 4 to 8 mm; for degree III crowding, the difference is more than 8 mm. The width of the crown is the distance between the most protruding points of the adjacent crown surfaces in the mesiodistal direction of the teeth, and the length of the existing arc of the dental arch is the distance from the mesial contact point of the first molar along the premolar surface to the cusp of the canine and then along the incisal edge of the maxillary incisor to the mesial contact point of the opposite first molar.

[0044] Deep overbite means that the midpoint of the incisal edge of the upper front teeth crown covers the midpoint of the incisal edge of the lower front teeth crown by more than one-third, and is divided into three degrees. In degree I deep overbite, the crown of the upper front teeth covers more than one-third of the crown of the lower front teeth but less than one-half; in degree II deep overbite, the crown of the upper front teeth covers more than one-half of the crown of the lower front teeth but less than two-thirds; in degree III deep overbite, the crown of the upper front teeth covers more than two-thirds of the crown of the lower front teeth.

[0045] Deep overjet means that the front-back distance between the midpoints of the incisors of the upper and lower front teeth exceeds 3mm, and is divided into three degrees: degree I deep overjet, with an overjet of 3-5mm; degree II deep overjet, with an overjet of 5-8mm; and degree III deep overjet, with an overjet of more than 8mm.

[0046] Open bite means that there is no overbite between the midpoints of the incisal edges of the upper and lower front teeth, and there is a gap in the vertical direction, which is divided into three degrees. In degree I open bite, the vertical gap between the midpoints of the incisal edges of the upper and lower front teeth is within 3mm; in degree II open bite, the vertical gap between the midpoints of the incisal edges of the upper and lower front teeth is between 3 and 5mm; in degree III open bite, the vertical gap between the midpoints of the incisal edges of the upper and lower front teeth is more than 5mm.

[0047] Malocclusion types are divided into three categories. The first type of malocclusion: neutral malocclusion, manifested by crowding of anterior teeth, protrusion of the upper dental arch, protrusion of both dental arches, crossbite of anterior teeth and buccolingual misalignment of posterior teeth. The second type of malocclusion: distal malocclusion, which is further divided into the first category, manifested by distal malocclusion and labial inclination of maxillary incisors; the first category subcategory, manifested by distal malocclusion on one side and neutral relationship on the other side; the second category, manifested by distal malocclusion and lingual inclination of maxillary incisors; the second category subcategory, manifested by distal malocclusion on one side and neutral relationship on the other side. The third type of malocclusion: mesial malocclusion, which is further divided into the third category subcategory, manifested by mesial malocclusion on one side and neutral malocclusion on the other side.

[0048] Step 104, constructing a causal model, the input of the causal model is the clinical diagnosis and treatment indications and industry diagnosis and treatment standards, and the output of the causal model is the treatment plan corresponding to the clinical diagnosis and treatment indications.

[0049] In this embodiment, the causal model can be constructed based on Bayesian networks, counterfactual reasoning, etc., and the relationship between input and output can be represented by a decision tree or a causal graph.

[0050] Step 105: training the multi-task learning model and the causal learning model based on the acquired clinical data and diagnosis and treatment data.

[0051] Step 106, obtaining orthodontic-related clinical data of the current patient, and after preprocessing the clinical data, inputting the trained multi-task learning model and causal model to obtain an auxiliary diagnosis and treatment plan, and generating a diagnosis and treatment report for the current patient.

[0052] like Figure 2 As shown, the diagnosis and treatment report includes patient information and auxiliary diagnosis and treatment suggestions, as well as at least one of the patient's facial photo analysis results, intraoral dental photo analysis results, dental image analysis results, and cephalometric analysis results, which are determined according to the orthodontic-related clinical data type input for the current patient. In addition, the diagnosis and treatment report may also include manually added feedback from professional dentists, which is used to dynamically optimize the multi-task learning model and the causal model.

[0053] Example 2

[0054] The above is an introduction to the method embodiment. The following is a further explanation of the solution of the present invention through a system embodiment.

[0055] This embodiment provides an artificial intelligence-based oral orthodontic treatment report generation system, which is used to implement the method described in the above embodiment 1. Figure 3 As shown, the system includes:

[0056] Data acquisition module 201: acquiring historical orthodontic related clinical data and diagnosis and treatment data corresponding to the historical orthodontic related clinical data;

[0057] Data preprocessing module 202: preprocessing the clinical data and diagnosis and treatment data;

[0058] Model construction module 203: includes a multi-task learning model construction submodule 2031 and a causal model construction submodule 2032, wherein the multi-task learning model construction submodule 2031 is used to construct a multi-task learning model based on artificial intelligence, the input of the multi-task learning model is multimodal data, including pre-processed clinical data and diagnosis and treatment data, and the output of the multi-task learning model is a plurality of clinical diagnosis and treatment indications, including one or more of tooth crowding, deep overbite, deep overjet, open bite, malocclusion type, skeletal and profile facial types; the causal model construction submodule 2032 is used to construct a causal model, the input of the causal model is the clinical diagnosis and treatment indications and industry diagnosis and treatment standards, and the output of the causal model is the treatment plan corresponding to the clinical diagnosis and treatment indications;

[0059] Model optimization module 204: training the multi-task learning model and causal learning model based on the acquired clinical data and diagnosis and treatment data;

[0060] Report generation module 205: includes a diagnosis report generation submodule 2051 and a treatment report generation submodule 2052, which are used to obtain the orthodontic-related clinical data of the current patient, and after pre-processing the clinical data, input the trained multi-task learning model and causal model. The diagnosis report generation submodule 2051 calls the output of the multi-task learning model to generate an auxiliary diagnosis report, and the treatment report generation submodule 2052 calls the output of the causal model to generate an auxiliary treatment plan, and generates a diagnosis and treatment report for the current patient based on the auxiliary diagnosis report and the auxiliary treatment plan.

[0061] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0062] Example 3

[0063] Figure 4 FIG. 1 is a schematic diagram of the structure of a terminal for generating an oral orthodontic treatment report in an embodiment of the present invention. Figure 4As shown, the report generating terminal 300 can be a computer system, including a memory 301, a processor 304 and a bus 306 connecting different system components. The memory 301 can be a system memory, a non-volatile storage medium, etc. The system memory can be a storage medium storing an operating system, an application program, a boot loader (Boot Loader) and other programs, etc., and can include a volatile storage medium, such as a random access memory (RAM) and / or a cache memory. The non-volatile storage medium, for example, stores instructions for executing at least one of the corresponding embodiments of the diagnostic report generating method. The non-volatile storage medium includes but is not limited to a disk memory, an optical memory, a flash memory, etc. The processor 301 can be implemented in a discrete hardware component manner such as a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistors. Accordingly, each module such as a data acquisition module, a data preprocessing module, and a generating module can be implemented by a central processing unit (CPU) running the instructions in the memory that execute the corresponding steps, or can be implemented by a dedicated circuit that executes the corresponding steps. Bus 306 may use any of a variety of bus architectures including, but not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MCA) bus, and a Peripheral Component Interconnect (PCI) bus.

[0064] In some embodiments, the report generation terminal 300 may further include an input / output interface 305, a network interface 302, a storage interface 303, etc. These interfaces may be connected to the memory 301 and the processor 304 via a bus 306. The input / output interface 305 may provide a connection interface for input / output devices such as a display, a mouse, and a keyboard. The network interface 302 may provide a connection interface for various networked devices. The storage interface 303 may provide a connection interface for external storage devices such as a floppy disk, a USB flash drive, and an SD card.

[0065] In some embodiments, the network interface 302 can be connected to a cloud server to obtain the patient's clinical information and other data at any time.

[0066] In summary, the present invention provides a method, system and terminal for generating diagnosis and treatment reports for orthodontics based on artificial intelligence. Based on the patient's orthodontic-related clinical data (multimodal data such as intraoral tooth data, facial features, medical history, etc.) and the clinical diagnosis and treatment records of professional orthodontists, by constructing a multi-task learning model and a causal model, the method automatically and accurately calculates the various parameters required in the orthodontic diagnosis and treatment process to generate an individualized auxiliary diagnosis report for the patient under the current conditions, and gives subsequent treatment plan recommendations; the method achieves the technical effect of using artificial intelligence technology to accurately identify orthodontic-related problems and generate diagnosis and treatment reports. The present invention can alleviate the problems of laborious diagnosis, inconsistent diagnostic standards, and poor diagnostic accuracy in the prior art, and effectively help clinical orthodontists to efficiently make diagnoses and determine treatment plans, thereby achieving the purpose of improving diagnosis and treatment efficiency.

[0067] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for generating an orthodontic diagnosis and treatment report based on artificial intelligence, characterized in that: The method comprises the following steps: Acquiring historical orthodontic related clinical data and preprocessing the clinical data; Acquiring diagnosis and treatment data corresponding to historical orthodontic-related clinical data, and preprocessing the diagnosis and treatment data; Constructing a multi-task learning model based on artificial intelligence, wherein the input of the multi-task learning model is multimodal data, including pre-processed clinical data and diagnosis and treatment data, and the output of the multi-task learning model is multiple clinical diagnosis and treatment indications, including one or more of dental crowding, deep overbite, deep overjet, open bite, malocclusion type, skeletal and profile facial type; Constructing a causal model, wherein the input of the causal model is the clinical diagnosis and treatment indication and the industry diagnosis and treatment standard, and the output of the causal model is the treatment plan corresponding to the clinical diagnosis and treatment indication; Training the multi-task learning model and the causal learning model based on the acquired clinical data and diagnosis and treatment data; The orthodontic-related clinical data of the current patient is obtained, and after preprocessing the clinical data, the trained multi-task learning model and causal model are input to obtain an auxiliary diagnosis and treatment plan, and a diagnosis and treatment report for the current patient is generated.

2. The method for generating an orthodontic treatment report based on artificial intelligence according to claim 1, characterized in that: The orthodontic-related clinical data include at least two of the following: intraoral dental images, facial photos, oral scan data, dental CBCT data, head positioning frontal and lateral radiographs, oral panoramic radiographs, temporomandibular magnetic resonance imaging data, and medical history data.

3. The method for generating an orthodontic diagnosis and treatment report based on artificial intelligence according to claim 2, characterized in that: The intraoral teeth pictures include at least two of the upper dentition images, lower dentition images, intraoral right side images, intraoral left side images, intraoral frontal images, facial front images, facial side images, and facial smile images taken by a digital camera.

4. The method for generating an orthodontic diagnosis and treatment report based on artificial intelligence according to claim 1, characterized in that: The multi-task learning model includes a multimodal feature extraction module and a multimodal feature fusion module. The multimodal feature extraction module uses a neural network to extract primary features and uses a large model to extract deep features. The multimodal feature fusion module fuses the features extracted by the multimodal feature extraction module according to the importance of the data modality.

5. The method for generating an orthodontic treatment report based on artificial intelligence according to claim 4, characterized in that: The importance of the data modality is set through weights based on clinical diagnosis and treatment experience.

6. The method for generating an orthodontic treatment report based on artificial intelligence according to claim 1, characterized in that: The preprocessing includes: denoising and standardizing facial photos, oral teeth pictures or other image pictures; and extracting key text information from medical history materials and diagnosis and treatment data using natural language processing.

7. The method for generating an orthodontic treatment report based on artificial intelligence according to claim 1, characterized in that: The diagnosis and treatment report includes patient information and auxiliary diagnosis and treatment suggestions, as well as at least one of the patient's facial photo analysis results, intraoral dental photo analysis results, dental image analysis results, and cephalometric analysis results, which is determined based on the input orthodontic-related clinical data type of the current patient.

8. The method for generating an orthodontic treatment report based on artificial intelligence according to claim 1, characterized in that: The diagnosis and treatment report includes manually added feedback from professional dentists, and the feedback is used to dynamically optimize the multi-task learning model and the causal model.

9. The method for generating an orthodontic treatment report based on artificial intelligence according to claim 1, characterized in that: The degree of crowding of the teeth indicates the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch, which is divided into three degrees. Degree I crowding indicates that the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch is 2 to 4 mm, degree II crowding indicates that the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch is 4 to 8 mm; degree III crowding indicates that the difference between the sum of the widths of the crowns and the length of the existing arc of the dental arch is more than 8 mm, wherein the crown width is the distance between the most protruding points of the adjacent crown surfaces in the mesiodistal direction of the teeth, and the length of the existing arc of the dental arch is the distance from the mesial contact point of the first molar along the premolar surface to the cusp of the canine and then along the incisal edge of the maxillary incisor to the mesial contact point of the first molar on the opposite side.

10. An artificial intelligence-based orthodontic diagnosis and treatment report generation system, characterized in that: For implementing the method according to any one of claims 1 to 9, the system comprises: Data acquisition module: acquiring historical orthodontic related clinical data and diagnosis and treatment data corresponding to the historical orthodontic related clinical data; Data preprocessing module: preprocessing the clinical data and diagnosis and treatment data; Model construction module: including a multi-task learning model construction submodule and a causal model construction submodule, wherein the multi-task learning model construction submodule is used to construct a multi-task learning model based on artificial intelligence, the input of the multi-task learning model is multimodal data, including pre-processed clinical data and diagnosis and treatment data, and the output of the multi-task learning model is a plurality of clinical diagnosis and treatment indications, including one or more of tooth crowding, deep overbite, deep overjet, open bite, malocclusion type, skeletal and profile facial types; the causal model construction submodule is used to construct a causal model, the input of the causal model is the clinical diagnosis and treatment indications and industry diagnosis and treatment standards, and the output of the causal model is the treatment plan corresponding to the clinical diagnosis and treatment indications; Model optimization module: training the multi-task learning model and causal learning model based on the acquired clinical data and diagnosis and treatment data; Report generation module: includes a diagnosis report generation submodule and a treatment report generation submodule, which are used to obtain the orthodontic-related clinical data of the current patient, and after pre-processing the clinical data, input the trained multi-task learning model and causal model. The diagnosis report generation submodule calls the output of the multi-task learning model to generate an auxiliary diagnosis report, and the treatment report generation submodule calls the output of the causal model to generate an auxiliary treatment plan, and generates a diagnosis and treatment report for the current patient based on the auxiliary diagnosis report and auxiliary treatment plan.

Images