Training data augmentation-based system for predicting facial soft tissue appearance after facial bone surgery

Abstract

This training data augmentation-based system for predicting facial soft tissue appearance after facial bone surgery comprises: a communication unit communicating with an external device; a user input unit into which a user's command is inputted; and a control unit configured to control the communication unit to collect CT image data of a patient showing soft tissue and bone through a network, learn the relationship between the soft tissue and the bone through artificial intelligence machine learning on the basis of the collected CT image data of the patient, and predict the appearance of a specific patient after surgery on the basis of a surgery plan of the specific patient and the learned relationship between the soft tissue and the bone when the CT image data of the specific patient and the surgery plan of the specific patient are inputted through the user input unit.

Description

Prediction system for facial soft tissue appearance after facial bone surgery based on expanded training data

[0001] The present invention relates to a system for predicting the external appearance of facial soft tissues after facial bone surgery based on expanded learning data, and more specifically, to a system for predicting the external appearance of facial soft tissues after facial bone surgery based on expanded learning data that uses artificial intelligence technology to predict and provide changes in facial soft tissues in the medical field, particularly after maxillofacial jaw correction surgery and facial contouring surgery.

[0002] Predicting the postoperative outcome of facial soft tissue appearance is crucial for correctly diagnosing conditions such as injuries, deformities, and tumors in the facial and oral regions, and for pursuing aesthetic restoration and functional recovery through surgical procedures and subsequent supportive treatments. Consultations are required to identify in advance changes in facial soft tissues resulting from bone alterations following orthognathic surgery or facial contouring surgery. To this end, the surgeon formulates a surgical plan by showing photographs of the patient's postoperative facial changes and explaining the procedure, or conducts simulations—that is, predictions—regarding orthognathic surgery, facial osteoplasty, and orthodontic treatment using imaging data and models of the patient's facial bones (jawbones).

[0003] Recently, as social health has become recognized as just as important as physical health, various cosmetic surgeries and orthodontic treatments—such as facial contouring, orthodontic treatment, and jaw correction (e.g., double jaw surgery)—are becoming popular. In particular, for surgeries that involve moving or reshaping facial bones, such as double jaw surgery or facial contouring, a meticulous surgical plan is required to predict changes in facial soft tissues resulting from bone movement.

[0004] Existing maxillofacial surgery prediction technologies have primarily focused on predicting pre- and post-operative soft tissue changes, but they faced the problem of difficulty in accurate prediction due to limitations in training data. Specifically, existing technologies were limited to data on pre- and post-operative changes and failed to adequately reflect the individual bone structure or soft tissue characteristics of each patient.

[0005] Therefore, the objective of the present invention is to provide a learning data expansion-based facial soft tissue appearance prediction system for post-operative facial bone surgery that can accurately predict the relationship between soft tissues and bone hard tissues by utilizing not only pre- and post-operative data but also various CT images as learning data, and based on this, predict patient-specific post-operative appearance changes.

[0006] A learning data expansion-based facial soft tissue shape prediction system for post-facial bone surgery according to the present invention, for achieving the above objective, comprises: a communication unit that communicates with an external device; a user input unit into which a user's command is input; and a control unit that controls the communication unit to collect CT image data of a patient showing soft tissue and bone through a network, and learns the relationship between soft tissue and bone through artificial intelligence machine learning on the collected CT image data of the patient, and when CT image data of a specific patient and a surgical plan of the specific patient are input through the user input unit, predicts the shape of the specific patient after surgery based on the surgical plan of the specific patient and the learned relationship between soft tissue and bone. After learning the relationship between soft tissue and bone, when CT image data of a specific patient and a surgical plan of the specific patient are input, the learned relationship between soft tissue and bone is utilized to precisely predict the relationship between soft tissue and hard bone tissue, and based on this, changes in the shape of the patient after surgery can be predicted in a patient-specific manner.

[0007] Here, the control unit comprises: a machine learning unit that learns the relationship between soft tissue and bone through artificial intelligence machine learning of collected CT image data of the patient; a skeletal structure generation unit that generates a skeletal structure after surgery based on the input surgical plan of the specific patient; a soft tissue generation unit that generates soft tissue after surgery using the relationship between soft tissue and bone learned by the machine learning unit based on the skeletal structure generated by the skeletal structure generation unit; and an image generation unit that generates an external image by synthesizing a two-dimensional face image of the patient with the soft tissue generated by the soft tissue generation unit. This is preferable because it generates a skeletal structure according to the surgical plan of the specific patient and generates soft tissue according to the skeletal structure generated using the learned relationship between soft tissue and bone, thereby generating an external image of the specific patient after surgery.

[0008] In addition, it is desirable to further include a display unit that displays the external image generated by the image generation unit, as this allows the patient to verify the post-operative external image generated for that specific patient.

[0009] Here, it is desirable for the control unit to further include a surgical planning support unit that supports surgical planning by providing the generated external image to the medical staff and the patient, as this can support smooth communication between the medical staff and the patient and the establishment of an optimal surgical plan.

[0010] Furthermore, it is desirable that the aforementioned machine learning unit, by learning pre- and post-operative data from various patients and the structural interactions between soft tissues, including skin and muscle, and bone, can utilize a wide range of data without being limited to specific surgical cases.

[0011] Here, it is desirable that the surgical plan for the specific patient includes information regarding which body part of the specific patient to amputate, in which direction to move it, and the depth and angle of the amputation, so that the postoperative skeletal structure can be modeled based on the surgical plan generated by reflecting the bone cutting position and degree of movement.

[0012] In addition, the above-mentioned skeletal structure generation unit is desirable because it can generate the skeletal structure after surgery more accurately by simulating the changes in the bone structure of the specific patient by applying the position, angle, and amount of movement of the bone after surgery using 3D modeling software.

[0013] Here, the soft tissue generation unit is desirable because the generated soft tissue data, including the thickness, location, and degree of deformation of the soft tissue after surgery, is generated to be output as a 3D model, allowing a specific patient to intuitively verify it in a report.

[0014] And the image generation unit integrates the 3D soft tissue model generated by the soft tissue generation unit with the 2D face image of the specific patient and converts it into an image for verifying external changes including the facial shape, contour, and skin condition of the specific patient after surgery to generate an external image. The generated external image is desirable because it can visually represent the external changes after surgery, allowing medical staff and the patient to easily understand them.

[0015] Here, the surgical planning support unit is desirable in that it provides visual materials including 3D images and 2D images that allow for comparison of external changes before and after surgery, so that external changes before and after surgery can be compared at once.

[0016] A method for predicting the external appearance of facial soft tissues after facial bone surgery based on learning data expansion according to the present invention for achieving the above objective comprises: a step of collecting CT image data of a patient in which soft tissues and bones are present through a network; a step of learning the relationship between soft tissues and bones through artificial intelligence machine learning on the collected CT image data of the patient; a step of inputting CT image data of a specific patient and a surgical plan of said specific patient; and a step of predicting the external appearance of said specific patient after surgery based on the surgical plan of said specific patient and the learned relationship between soft tissues and bones. After learning the relationship between soft tissues and bones, when the CT image data of a specific patient and the surgical plan of said specific patient are input, the relationship between soft tissues and bone hard tissues is precisely predicted by utilizing the learned relationship between soft tissues and bones, and based on this, changes in the external appearance after surgery tailored to the patient can be predicted.

[0017] Here, the step of predicting the external appearance of the specific patient comprises: a step of generating a postoperative skeletal structure based on the input surgical plan of the specific patient; a step of generating postoperative soft tissue using the learned relationship between soft tissue and bone based on the generated skeletal structure; and, by synthesizing a two-dimensional face image of the patient with the generated soft tissue to create an external appearance image, a skeletal structure is generated according to the surgical plan of the specific patient, and soft tissue is generated according to the skeletal structure generated using the learned relationship between soft tissue and bone, thereby generating the postoperative external appearance image of the specific patient.

[0018] Furthermore, the step of predicting the specific patient's appearance described above is desirable if it further includes a step of displaying the generated appearance image, as this allows the specific patient to verify the generated post-operative appearance image of the specific patient.

[0019] Here, it is desirable to include an additional step of providing the generated external images to the medical team and the patient to support surgical planning, as this facilitates smooth communication between the medical team and the patient and supports the establishment of an optimal surgical plan.

[0020] Furthermore, the step of learning the relationship between the soft tissue and bone is desirable because it allows for the utilization of a wide range of data without being limited to specific surgical cases by learning pre- and post-operative data from various patients and the structural interactions between soft tissues, including skin and muscle, and bone.

[0021] Here, it is desirable that the surgical plan for the specific patient includes information regarding which body part of the specific patient to amputate, in which direction to move it, and the depth and angle of the amputation, so that the postoperative skeletal structure can be modeled based on the surgical plan generated by reflecting the bone cutting position and degree of movement.

[0022] And the step of generating the skeletal structure after surgery is desirable if it includes a step of simulating changes in the bone structure of the specific patient by applying the position, angle, and amount of movement of the bone after surgery using 3D modeling software, so that the skeletal structure after surgery can be generated more accurately.

[0023] Here, the step of generating soft tissue after surgery is desirable because the generated soft tissue data, including the thickness, location, and degree of deformation of the soft tissue after surgery, is generated to be output as a 3D model, allowing a specific patient to intuitively verify it in a report.

[0024] And the step of generating the above external image is desirable in that it includes the step of integrating the generated 3D soft tissue model with the 2D face image of the specific patient; and converting it into an image for verifying external changes including the facial shape, contour, and skin condition of the specific patient after surgery. The generated external image can visually represent the external changes after surgery, making it easy for medical staff and patients to understand.

[0025] Here, the step of supporting the surgical plan is desirable if visual materials including 3D images and 2D images that allow comparison of external changes before and after surgery are provided, so that external changes before and after surgery can be compared at once.

[0026] According to the present invention, after learning the relationship between soft tissue and bone, when CT image data and a specific patient's surgical plan are input, the learned relationship between soft tissue and bone is utilized to precisely predict the relationship between soft tissue and hard bone tissue, and based on this, there is an effect of predicting changes in the external appearance after surgery tailored to the patient.

[0027] In addition, it has the effect of generating a skeletal structure according to a specific patient's surgical plan and generating soft tissues based on the generated skeletal structure using the learned relationship between soft tissues and bones, thereby enabling the creation of a specific patient's external appearance image after surgery.

[0028] In addition, it has the effect of allowing verification of a specific patient's post-operative external image generated for that specific patient.

[0029] In addition, it is effective in supporting smooth communication between medical staff and patients and establishing an optimal surgical plan.

[0030] In addition, it has the effect of enabling the utilization of a wide range of data, rather than being limited to specific surgical cases.

[0031] In addition, it has the effect of modeling the postoperative skeletal structure based on a surgical plan generated by reflecting the bone cutting position and degree of displacement.

[0032] FIG. 1 is a control block diagram of a facial soft tissue shape prediction system based on learning data expansion after facial bone surgery according to the present invention.

[0033] Figures 2 to 5 are 3D modeling examples of a facial soft tissue shape prediction system based on expanded learning data after facial bone surgery.

[0034] FIG. 6 is a flowchart of a first embodiment of a method for predicting the shape of facial soft tissue after facial bone surgery based on learning data expansion according to the present invention.

[0035] Figure 7 is a flowchart of a second embodiment of a method for predicting the shape of facial soft tissue after facial bone surgery based on expanded learning data.

[0036] Figure 8 is a flowchart of a third embodiment of a method for predicting the shape of facial soft tissue after facial bone surgery based on expanded learning data.

[0037] Figure 9 is a flowchart of the fourth embodiment of a facial soft tissue shape prediction system based on expanded learning data after facial bone surgery.

[0038] Hereinafter, a learning data expansion-based facial soft tissue shape prediction system (1) according to a preferred embodiment of the present invention will be described in detail with reference to the attached drawings.

[0039] FIG. 1 is a control block diagram of a facial soft tissue shape prediction system (1) based on learning data expansion after facial bone surgery according to the present invention.

[0040] A facial soft tissue shape prediction system (1) based on expanded learning data after facial bone surgery includes a communication unit (10), a user input unit (20), a display unit (30), a camera (40), and a control unit (50).

[0041] The communication unit (10) communicates with an external device. The communication unit (10) can perform wireless communication, and the wireless communication includes at least one of infrared communication, RF, Zigbee, and Bluetooth. The communication unit (10) receives a video signal and transmits it to the control unit (50) to be described later, and can be implemented in various ways corresponding to the specifications of the received video signal and the implementation form of the user terminal. For example, the communication unit (10) can wirelessly receive an RF (radio frequency) signal transmitted from a broadcasting station (not shown), or receive a video signal according to composite video, component video, super video, SCART, HDMI (high definition multimedia interface) specifications, etc. via a wired connection. If the video signal is a broadcast signal, the communication unit (10) may include a tuner that tunes the broadcast signal by channel.

[0042] The user input unit (20) receives commands from the user. The user input unit (20) can receive touch input from the user or remote input from the user using a remote controller and transmit it to the corresponding control unit (50). Additionally, the user input unit (20) can receive voice input spoken by the user and transmit the voice signal to the control unit (50). In that case, the user input unit (20) can be implemented, for example, as a microphone. The user input unit (20) can also perform signal processing on the received voice signal itself. However, the form of user input that the user input unit (20) can receive is not limited to this, and, for example, user input through motion recognition, etc., can also be received.

[0043] The display unit (30) can display an external image generated by the image generation unit (54). The display unit (30) displays an image based on an image signal processed by image processing. The implementation method of the display unit (30) is not limited and can be implemented in various display methods such as liquid crystal, plasma, light-emitting diode, organic light-emitting diode, surface conduction electron-emitter, carbon nanotube, nanocrystal, etc.

[0044] The display unit (30) may additionally include additional configurations depending on the implementation method. For example, if the display unit (30) is a liquid crystal type, the display unit (30) includes a liquid crystal display panel (not shown), a backlight unit (not shown) that supplies light thereto, and a panel driving board (not shown) that drives the panel (not shown). The display unit (30) may display a voice recognition result as information regarding the recognized voice. Here, the voice recognition result can be displayed in various forms such as text, graphics, and icons, and the text includes characters and numbers. The display unit (30) may further display candidate commands and application information based on the voice recognition result. The user can check whether the voice has been correctly recognized by the voice recognition result displayed on the display unit (30), and can select a command corresponding to the voice spoken by the user from among the displayed candidate commands by operating the user input unit (30) provided on the remote control, or select and check information related to the voice recognition result.

[0045] The camera (40) can capture the outside and the front, including the user.

[0046] The control unit (50) controls the communication unit (10) to collect CT image data of a patient showing soft tissue and bone through a network, and learns the relationship between soft tissue and bone through artificial intelligence machine learning on the collected CT image data of the patient, and when CT image data of a specific patient and a surgical plan of a specific patient are input through the user input unit (20), the external appearance of the specific patient after surgery is predicted based on the surgical plan of the specific patient and the learned relationship between soft tissue and bone.

[0047] A surgical plan for a specific patient may include information regarding which body part of the patient to amputate, in which direction to move it, and the depth and angle of the amputation.

[0048] The control unit (50) may further include a surgical planning support unit (55) that supports surgical planning by providing the generated external image to medical staff and patients.

[0049] The control unit (50) includes a machine learning unit (51), a skeletal structure generation unit (52), a soft tissue generation unit (53), an image generation unit (54), and a surgical planning support unit (55).

[0050] The machine learning unit (51) can learn the relationship between soft tissue and bone through artificial intelligence machine learning on collected patient CT image data. The machine learning unit (51) can learn various patient pre- and post-operative data, and the structural interaction between soft tissue and bone, including skin and muscle.

[0051] The skeletal structure generation unit (52) can generate a post-operative skeletal structure based on the input surgical plan of a specific patient. The skeletal structure generation unit (52) can simulate changes in the bone structure of a specific patient by applying the position, angle, and amount of movement of the bone after surgery using 3D modeling software.

[0052] The soft tissue generation unit (53) can generate postoperative soft tissue by utilizing the relationship between soft tissue and bone learned by the machine learning unit (51) based on the skeletal structure generated by the skeletal structure generation unit (52). The soft tissue generation unit (53) can generate soft tissue data including the thickness, location, and degree of deformation of the postoperative soft tissue so that it is output as a 3D model.

[0053] The image generation unit (54) can generate an external image by synthesizing a patient's 2D face image with the soft tissue generated by the soft tissue generation unit (53). The image generation unit (54) can generate an external image by integrating a 3D soft tissue model generated by the soft tissue generation unit (53) with a specific patient's 2D face image and converting it into an image to check external changes including the specific patient's face shape, contour, and skin condition after surgery.

[0054] The surgery planning support department (55) can provide visual materials including 3D images and 2D images that allow comparison of external changes before and after surgery.

[0055] FIG. 6 is a flowchart of a first embodiment of a method for predicting the shape of facial soft tissue after facial bone surgery based on learning data expansion according to the present invention.

[0056] CT image data of a patient showing soft tissue and bone is collected through a network (S1).

[0057] The relationship between soft tissue and bone is learned through artificial intelligence machine learning using collected CT image data of the patient (S2).

[0058] CT image data of a specific patient and a surgical plan for a specific patient are entered (S3).

[0059] Predicts the appearance of a specific patient after surgery based on the surgical plan and the learned relationship between soft tissue and bone (S4).

[0060] Figure 7 is a flowchart of a second embodiment of a method for predicting the shape of facial soft tissue after facial bone surgery based on expanded learning data.

[0061] CT image data of a patient showing soft tissue and bone is collected through a network (S11).

[0062] The relationship between soft tissue and bone is learned through artificial intelligence machine learning using collected CT image data of the patient (S12).

[0063] CT image data of a specific patient and a surgical plan for a specific patient are entered (S13).

[0064] Based on the input surgical plan for a specific patient, a postoperative skeletal structure is generated (S14).

[0065] Postoperative soft tissue is generated using the relationship between soft tissue and bone learned based on the generated skeletal structure (S15).

[0066] An external image is generated by synthesizing the patient's 2D face image onto the generated soft tissue (S16).

[0067] Displays the generated external image (S17).

[0068] Figure 8 is a flowchart of a third embodiment of a method for predicting the shape of facial soft tissue after facial bone surgery based on expanded learning data.

[0069] CT image data of a patient showing soft tissue and bone is collected through a network (S21).

[0070] Artificial intelligence machine learning learns pre- and post-operative data of various patients, as well as the structural interactions between soft tissues including skin and muscle and bone (S22).

[0071] A surgical plan for a specific patient is input, which includes CT image data of the specific patient and information on which body part of the specific patient will be amputated, in which direction it will be moved, and the depth and angle of the amputation (S23).

[0072] Based on the input surgical plan for a specific patient, 3D modeling software is used to simulate changes in the bone structure of the specific patient by applying the position, angle, and amount of movement of the bone after surgery, thereby generating a skeletal structure after surgery (S24).

[0073] Based on the generated skeletal structure, the relationship between the soft tissue and the bone learned is used to generate the soft tissue after surgery, including the thickness, location, and degree of deformation of the soft tissue after surgery, so that the generated soft tissue data is output as a 3D model (S25).

[0074] The generated 3D soft tissue model is integrated with the 2D face image of the specific patient (S26).

[0075] After surgery, an external image is generated by converting it into an image to verify external changes including the facial shape, contour, and skin condition of the specific patient (S27).

[0076] Displays the generated external image (S28).

[0077] Figure 9 is a flowchart of a fourth embodiment of a method for predicting the shape of facial soft tissue after facial bone surgery based on expanded learning data.

[0078] CT image data of a patient showing soft tissue and bone is collected through a network (S31).

[0079] Artificial intelligence machine learning learns pre- and post-operative data of various patients, as well as the structural interactions between soft tissues including skin and muscle and bone (S32).

[0080] A surgical plan for a specific patient is input, which includes CT image data of the specific patient and information on which body part of the specific patient will be amputated, in which direction it will be moved, and the depth and angle of the amputation (S33).

[0081] Based on the input surgical plan for a specific patient, 3D modeling software is used to simulate changes in the bone structure of the specific patient by applying the position, angle, and amount of movement of the bone after surgery, thereby generating a skeletal structure after surgery (S34).

[0082] Based on the generated skeletal structure, the relationship between the soft tissue and the bone learned is used to generate the soft tissue after surgery, including the thickness, location, and degree of deformation of the soft tissue after surgery, so that the generated soft tissue data is output as a 3D model (S35).

[0083] The generated 3D soft tissue model is integrated with the 2D face image of the specific patient (S36).

[0084] After surgery, an external image is generated by converting it into an image to verify external changes including the facial shape, contour, and skin condition of the specific patient (S37).

[0085] Displays the generated external image (S38).

[0086] Visual materials including 3D and 2D images that allow comparison of external changes before and after surgery are provided to the medical staff and patient to support surgical planning (S39).

[0087] Here, the technical means of the facial soft tissue shape prediction system (1) based on expanded learning data after facial bone surgery is described.

[0088] Various CT image data are collected, and based on this, the relationship between soft tissues and bone hard tissues is learned using machine learning technology.

[0089] It receives a specific patient's preoperative CT image and models the postoperative skeletal structure based on the patient's surgical plan (location and extent of bone resection).

[0090] Changes in soft tissue are precisely generated by applying learned data and algorithms to the modeled postoperative skeletal structure.

[0091] The generated soft tissue changes are combined with the patient's 2D facial image to create an intuitively understandable generated image, and the surgical plan can be adjusted based on this.

[0092] As a result, by expanding the range of training data compared to existing technologies, it is possible to predict soft tissue more accurately and precisely.

[0093] By visualizing postoperative soft tissue changes in a patient-specific manner, it supports smooth communication between medical staff and patients and the establishment of an optimal surgical plan.

[0094] Even when no soft tissue remains, soft tissue characteristics and external appearance can be predicted based on the skeletal structure, thereby securing differentiation from existing technologies.

[0095] By combining machine learning and visualization technologies, it is possible to specifically predict the patient's post-operative appearance.

[0096] Here, the implementation details of the facial soft tissue shape prediction system (1) based on expanded learning data after facial bone surgery are explained.

[0097] CT images from various patients are collected to learn the relationship between soft tissues and bone / hard tissues. This enables the utilization of a wide range of data, rather than being limited to specific surgical cases.

[0098] It receives CT images of a patient scheduled for surgery, analyzes the patient's skeletal structure, and specifically establishes the surgical plan (location and extent of amputation).

[0099] The postoperative skeletal structure is modeled based on the input surgical plan. This model is generated by reflecting the bone resection location and the degree of displacement.

[0100] The modeled skeletal structure is compared with machine-learned data to generate the expected shape of the soft tissue after surgery. This generation is performed based on the relationships between soft tissue and bone learned from various patient data.

[0101] An intuitive image is generated by synthesizing the patient's 2D facial image with the generated soft tissue changes. This generated image visually represents the changes in appearance after surgery, allowing medical staff and the patient to easily understand them.

[0102] The generated results are provided to the medical team and the patient to enable the optimization of the surgical plan, such as adjusting the resection location or amount.

[0103] Due to the above-mentioned facial soft tissue shape prediction system (1) and the post-operative shape prediction method based on expanded learning data, when CT image data of a specific patient and a surgical plan of a specific patient are input after learning the relationship between soft tissue and bone, the relationship between soft tissue and bone hard tissue can be precisely predicted by utilizing the learned relationship between soft tissue and bone hard tissue, and based on this, the patient-specific post-operative shape change can be predicted.

[0104] In addition, a skeletal structure is generated according to a specific patient's surgical plan, and soft tissues corresponding to the generated skeletal structure are created using the learned relationship between soft tissues and bones, thereby generating an external image of the specific patient after surgery.

[0105] In addition, it is possible to check the post-operative external appearance image of a specific patient generated for that patient.

[0106] In addition, it can support smooth communication between medical staff and patients and the establishment of an optimal surgical plan.

[0107] In addition, it is possible to utilize a wide range of data without being limited to specific surgical cases.

[0108] In addition, the postoperative skeletal structure can be modeled based on a surgical plan generated by reflecting the bone cutting position and degree of displacement.

[0109] The facial soft tissue shape prediction system based on the learning data expansion of the present invention has industrial applicability because, after learning the relationship between soft tissue and bone, it utilizes the learned relationship between soft tissue and bone when a specific patient's CT image data and surgical plan are input, thereby precisely predicting the relationship between soft tissue and hard bone tissue, and based on this, it can predict patient-specific postoperative shape changes.

Claims

1. In a system for predicting the external appearance of facial soft tissues after facial bone surgery based on the expansion of training data, A communication unit that communicates with an external device; A user input section into which a user's command is entered; and A learning data expansion-based facial soft tissue appearance prediction system after facial bone surgery, characterized by including a control unit that controls the communication unit to collect CT image data of a patient in which soft tissue and bone are present through a network, learns the relationship between soft tissue and bone through artificial intelligence machine learning of the collected CT image data of the patient, and predicts the appearance of the specific patient after surgery based on the surgical plan of the specific patient and the learned relationship between soft tissue and bone when CT image data of the specific patient and the surgical plan of the specific patient are input through the user input unit.

2. In Paragraph 1, The above control unit is, A machine learning unit that enables the collected patient CT image data to learn the relationship between soft tissue and bone through artificial intelligence machine learning; A skeletal structure generation unit that generates a postoperative skeletal structure based on the input surgical plan of the specific patient; and A soft tissue generation unit that generates postoperative soft tissue using the relationship between soft tissue and bone learned by the machine learning unit based on the skeletal structure generated by the skeletal structure generation unit; and A learning data expansion-based facial soft tissue shape prediction system after facial bone surgery, characterized by including an image generation unit that generates an external shape image by synthesizing a patient's two-dimensional face image with the soft tissue generated by the soft tissue generation unit.

3. In Paragraph 2, A learning data expansion-based facial soft tissue shape prediction system after facial bone surgery, characterized by further including a display unit that displays an external shape image generated by the image generation unit.

4. In Paragraph 2, The above control unit is, A learning data expansion-based facial soft tissue shape prediction system after facial bone surgery, characterized by further including a surgical planning support unit that provides generated shape images to medical staff and patients to support surgical planning.

5. In Paragraph 2, The above machine learning unit is, A learning data expansion-based facial soft tissue shape prediction system for post-facial bone surgery, characterized by learning pre- and post-operative data from various patients and the structural interaction between soft tissues, including skin and muscle, and bone.

6. In Paragraph 2, The surgical plan for the specific patient mentioned above is, A learning data expansion-based facial soft tissue shape prediction system after facial bone surgery, characterized by including information on which body part of the specific patient to cut, in which direction to move, and the depth and angle of the cut.

7. In Paragraph 6, The above-mentioned skeletal structure generating unit is, A learning data expansion-based facial soft tissue shape prediction system after facial bone surgery, characterized by simulating changes in the bone structure of the specific patient by applying the position, angle, and displacement amount of the bone after surgery using 3D modeling software.

8. In Paragraph 2, The above soft tissue generating unit is, A learning data expansion-based facial soft tissue shape prediction system after facial bone surgery, characterized in that generated soft tissue data including the thickness, location, and degree of deformation of the soft tissue after surgery is generated to be output as a 3D model.

9. In Paragraph 2, The above image generation unit is, A learning data expansion-based facial soft tissue appearance prediction system after facial bone surgery, characterized by integrating a 3D soft tissue model generated in the soft tissue generation unit with a 2D facial image of the specific patient, and converting it into an image for verifying external changes including the facial shape, contour, and skin condition of the specific patient after surgery to generate an external image.

10. In Paragraph 4, The aforementioned surgical planning support department, A learning data expansion-based facial soft tissue shape prediction system after facial bone surgery, characterized by providing visual materials including 3D and 2D images that allow comparison of changes in appearance before and after surgery.

11. In a method for predicting the external appearance of facial soft tissues after facial bone surgery based on the expansion of training data, A step of collecting CT image data of a patient showing soft tissue and bone through a network; A step of learning the relationship between soft tissue and bone through artificial intelligence machine learning using collected patient CT image data; A step in which CT image data of a specific patient and a surgical plan for the specific patient are input; A method for predicting the appearance of facial soft tissue after facial bone surgery based on expanded learning data, characterized by including a step of predicting the appearance of a specific patient after surgery based on the surgical plan of the specific patient and the learned relationship between soft tissue and bone.

12. In Paragraph 11, The step of predicting the appearance of the specific patient mentioned above is, A step of generating a postoperative skeletal structure based on the input surgical plan of the specific patient; and A step of generating postoperative soft tissue using the relationship between soft tissue and bone learned based on the generated skeletal structure; and A method for predicting the external shape of facial soft tissue after facial bone surgery based on expanded learning data, characterized by including the step of generating an external shape image by synthesizing a patient's two-dimensional face image with the generated soft tissue.

13. In Paragraph 12, The step of predicting the appearance of the specific patient mentioned above is, A method for predicting the shape of facial soft tissue after facial bone surgery based on expanded learning data, characterized by further including a step of displaying a generated shape image.

14. In Paragraph 12, A method for predicting the shape of facial soft tissue after facial bone surgery based on expanded learning data, characterized by including an additional step of providing generated shape images to medical staff and patients to support surgical planning.

15. In Paragraph 11, The step of learning the relationship between the soft tissue and bone mentioned above is, A method for predicting the appearance of facial soft tissues after facial bone surgery based on expanded learning data, characterized by including a step of learning pre- and post-operative data of various patients and the structural interaction between soft tissues, including skin and muscle, and bone.

16. In Paragraph 12, The surgical plan for the specific patient mentioned above is, A method for predicting the external shape of facial soft tissue after facial bone surgery based on expanded learning data, characterized by including information on which body part of the specific patient to cut, in which direction to move, and the depth and angle of the cut.

17. In Paragraph 16, The step of generating the skeletal structure after the above surgery is, A method for predicting the external appearance of facial soft tissue after facial bone surgery based on expanded learning data, characterized by including a step of simulating changes in the bone structure of a specific patient by applying the position, angle, and displacement amount of the bone after surgery using 3D modeling software.

18. In Paragraph 12, The step of generating soft tissue after the above surgery is, A method for predicting the external appearance of facial soft tissue after facial bone surgery based on training data expansion, characterized by including a step of generating soft tissue data, including the thickness, location, and degree of deformation of the soft tissue after surgery, to be output as a 3D model.

19. In Paragraph 12, The step of generating the above external image is, A step of integrating the generated 3D soft tissue model with the 2D face image of the specific patient; and A method for predicting the external appearance of facial soft tissue after facial bone surgery based on expanded learning data, characterized by including the step of generating an external image by converting the data into an image to verify external changes including the facial shape, contour, and skin condition of the specific patient after surgery.

20. In Paragraph 14, The steps supporting the above surgical plan are, A method for predicting the appearance of facial soft tissue after facial bone surgery based on expanded learning data, characterized by including a step of providing visual materials including 3D images and 2D images that allow comparison of changes in appearance before and after surgery.

Images