TWI756681B - Artificial intelligence assisted evaluation method applied to aesthetic medicine and system using the same - Google Patents

Abstract

The present invention discloses an artificial intelligence assisted evaluation method applied to aesthetic medicine, which is applied to an artificial intelligence identified and assisted module. Among this, a real-time facial expression evaluation result regarding a testee is inputted to the artificial intelligence identified and assisted module. Then at least one of a medical knowledge rule module and a historical database of aesthetic medical auxiliary evaluation results are selected for executing an artificial intelligence identified and assisted procedure. After that, the artificial intelligence identified and assisted module generates and outputs an aesthetic medical auxiliary evaluation result. Thus, a therapeutic effect for personalized beauty can be acquired via performing an aesthetic medical behavior based on the aesthetic medical auxiliary evaluation result.

Description

Artificial intelligence-assisted evaluation method and system for cosmetic medicine

The present invention relates to an artificial intelligence (Artificial Intelligence, AI) assisted evaluation method, especially an artificial intelligence assisted evaluation method based on facial expressions and applied to beauty medicine and a system using the method.

Today's popular cosmetic medical treatment, especially the facial micro-plastic cosmetic treatment, has been widely loved and accepted by people of all ages.

Among them, the treatment methods of facial micro-surgery and beauty treatment mainly rely on the medical professional skills and knowledge of the physicians themselves, and handle them with general or normalized standard procedures, or add some individual physicians’ own practice experience. Judgment; however, these conventional methods usually lack in-depth customization considerations, so that there is often a more or less expectation gap between the actual cosmetic treatment results and the pre-expected treatment effects.

What's more, some of the actual treatment results after cosmetic medical treatment, due to the wrong individual judgment of the physicians themselves, lead to worse treatment results after surgery. As a result, more medical treatment disputes and lack of medical treatment are caused.

Therefore, how to propose more and better methods and tools to assist in the treatment of individualized cosmetic medical needs is a technical issue to be solved by the cosmetic medical industry at present.

The main purpose of the present invention is to provide an artificial intelligence (AI) assisted evaluation method for cosmetic medical treatment based on facial expressions and a system using the method, which can meet the needs of individualized cosmetic medical treatment, so as to solve the above-mentioned conventional wisdom. Lack of technology.

The implementation concept of this case is mainly to use the facial expression evaluation module to first conduct a personalized real-time facial expression evaluation result. Combined with the functional medical anatomy rules and dynamic medical anatomy rules in the medical knowledge rules module and the historical database of cosmetic medical assistant assessment results, the results of cosmetic medical assistant assessment can be provided, so as to provide exclusive and personalized cosmetic medical assistant suggestions.

In order to achieve the above object, in the first preferred embodiment, the present invention provides an artificial intelligence-assisted evaluation method applied in cosmetic medicine, which at least includes the following steps: providing a real-time facial expression evaluation result of a subject; inputting the The real-time facial expression evaluation result is put into an artificial intelligence identification analysis module; and at least one of a medical knowledge rule module and a beauty medical auxiliary evaluation result historical database is selected and matched to execute an artificial intelligence identification analysis program, And generate and output a cosmetic medical assistant evaluation result.

Preferably, it further includes the following steps: feedback and store the cosmetic The medical assistant evaluation result is sent to at least one of the medical knowledge rule module and the aesthetic medical assistant evaluation result historical database.

Preferably, the output of the cosmetic medical assistant evaluation result at least includes: the combination and the preferred order of the results of the evaluation of the treatment site by one of the subjects, or the combination and the preferred order of the results of the evaluation of the treatment site, and an injection filler type. with dosage.

Preferably, the medical knowledge rule module further includes a functional medical anatomy rule and a dynamic medical anatomy rule.

Preferably, the real-time facial expression evaluation result includes: a static expression evaluation result, or the static expression evaluation result and a dynamic expression evaluation result.

Preferably, providing the real-time facial expression evaluation result further includes the following steps: classifying a face into plural facial action units according to the medical knowledge rule module; At least one of the results is used to form a plurality of emotion index combinations; and the real-time facial expression evaluation result is formed according to the proportional result of each emotion index combination.

Preferably, the sentiment index combinations are at least one of a positive sentiment index combination and a negative sentiment index combination.

Preferably, the combination of negative emotion indicators is one of a sadness indicator, an anger indicator, a worry indicator, a surprise indicator, a fear indicator, a contempt indicator, and a dislike indicator, or a combination thereof.

Preferably, the combination of positive emotion indicators is one of a happiness indicator, a happiness indicator, a satisfaction indicator, a moving indicator, a positive indicator, and a relaxation indicator, or a combination thereof.

Preferably, the historical database of cosmetic medical assistant evaluation results includes plural cosmetic medical assistant evaluation results, which are used for the artificial intelligence identification and analysis module to execute an artificial intelligence according to at least one artificial intelligence deep learning/training algorithm. Deep learning/training procedures.

Preferably, wherein, the artificial intelligence deep learning/training program includes the following steps: providing the cosmetic medical assistant evaluation results; wherein, each of the cosmetic medical assistant evaluation results at least includes: a historical subject-related basic data , a person's facial expression evaluation results, a person's facial features, a functional medical anatomy rule and a dynamic medical anatomy rule in the medical knowledge rule module, a combination and priority order of the results of evaluating the treatment site, and an injection filler types and doses; and inputting the aesthetic medical assistant evaluation results into the artificial intelligence identification and analysis module.

Preferably, the personal facial feature includes a static texture feature of a habitual expression, a static contour line feature or a skin texture feature.

Preferably, the at least one artificial intelligence deep learning/training algorithm is at least one of an artificial neural network road algorithm and a deep learning algorithm.

In a second preferred embodiment, the present invention provides an electronic device using the aforementioned artificial intelligence-assisted evaluation method, the electronic device at least includes: a facial expression evaluation module for providing the real-time facial expression evaluation result; The artificial intelligence identification and analysis module includes the artificial intelligence identification and analysis program, the artificial intelligence identification and analysis program receives the real-time facial expression evaluation result, and generates the beauty medical assistant evaluation result; and an input and output module, which outputs the beauty treatment A medical assistant evaluation result; wherein, the artificial intelligence recognition and analysis module receives at least one human facial feature from at least one of the facial expression evaluation module and the input and output module.

Preferably, the electronic device is connected to at least one of the historical database of aesthetic medical aid evaluation results and the medical knowledge rule module by at least one of a wireless transmission method and a wired transmission method.

Preferably, the electronic device is a handheld smart mobile device, a personal computer (PC) or a stand-alone smart device.

In a third preferred embodiment, the present invention provides an artificial intelligence-assisted evaluation system for cosmetic medicine, which at least includes: a facial expression evaluation module for providing a real-time facial expression evaluation result of a subject; and an artificial intelligence The identification and analysis module is connected to the facial expression evaluation module; wherein, the artificial intelligence identification and analysis module receives the real-time facial expression evaluation result, and evaluates according to a connected medical knowledge rule module and a cosmetic medical aid At least one of the result history databases is used to execute an artificial intelligence identification and analysis program, and to adaptively generate and output a cosmetic medical assistant evaluation result.

Preferably, the artificial intelligence identification and analysis module feeds back and stores the cosmetic assistance evaluation result to at least one of the medical knowledge rule module and the historical database of cosmetic medical assistance evaluation results.

Preferably, the cosmetic medical assistant evaluation result includes at least the combination and the preferred order of the results of one of the subjects evaluating the treatment site, or the combination and the preferred order of the evaluation results of the treatment site, and an injection filler type and dosage.

Preferably, the medical knowledge rule module further includes a functional medical anatomy rule and a dynamic medical anatomy rule.

Preferably, the real-time facial expression evaluation result includes a static expression evaluation result, or the static expression evaluation result and a dynamic expression evaluation result.

Preferably, the facial expression evaluation module includes: a facial image capture unit for performing an image capture action to obtain a real-time facial image; and a facial motion encoding unit, according to the real-time The facial image and the medical knowledge rule module distinguish a real-time facial action presented in the image into a plurality of facial action units; wherein, according to a detection result of each of the facial action units and at least another of the facial action units The change between another detection result of the facial action unit forms the static expression evaluation result and the dynamic expression evaluation result.

Preferably, at least one of the facial expression evaluation module and the artificial intelligence recognition analysis module further includes: an emotion analysis and facial recognition unit, based on the static expression evaluation result and the dynamic expression evaluation result At least one of the emotion index combinations is formed; wherein, the real-time facial expression evaluation result is formed according to the proportional result of each of the emotion index combinations.

Preferably, the sentiment index combinations are at least one of a positive sentiment index combination and a negative sentiment index combination.

Preferably, the combination of negative emotion indicators is one of a sadness indicator, an anger indicator, a worry indicator, a surprise indicator, a fear indicator, a contempt indicator, and a dislike indicator, or a combination thereof.

Preferably, the combination of positive emotion indicators is one of a happiness indicator, a happiness indicator, a satisfaction indicator, a moving indicator, a positive indicator, and a relaxation indicator, or a combination thereof.

Preferably, the historical database of cosmetic medical assistant evaluation results includes plural cosmetic medical assistant evaluation results, which are used for the artificial intelligence identification and analysis module to execute an artificial intelligence according to at least one artificial intelligence deep learning/training algorithm. Deep learning/training procedures.

Preferably, wherein, each of the cosmetic medical assistant evaluation results to At least include: a basic information related to a historical subject, a person's facial expression evaluation result, a person's facial features, a functional medical anatomy rule in the medical knowledge rule module, a dynamic medical anatomy rule, an evaluation treatment The combination and preferred order of site results and one or a combination of the type and dose of an injectable filler.

Preferably, the personal facial feature includes a static texture feature of a habitual expression, a static contour line feature or a skin texture feature.

Preferably, the personal facial features are provided to the artificial intelligence recognition and analysis module from at least one of the facial expression evaluation module and an input and output module.

Preferably, the artificial intelligence deep learning/training program is to input the aesthetic medical assistant evaluation results to the artificial intelligence identification and analysis module.

Preferably, the at least one artificial intelligence deep learning/training algorithm is at least one of an artificial neural network road algorithm and a deep learning algorithm.

Preferably, it is assembled to form an electronic device by the facial expression evaluation module, an input and output module and the artificial intelligence recognition and analysis module; wherein, the electronic device is a handheld smart mobile device, a person A computer (PC) or a stand-alone smart device.

Preferably, the electronic device is connected to at least one of the historical database of aesthetic medical aid evaluation results and the medical knowledge rule module by at least one of a wireless transmission method and a wired transmission method.

100: Artificial Intelligence Assisted Evaluation System

110: Facial Expression Evaluation Module

111: Facial Action Coding Unit

112: Sentiment Analysis and Facial Recognition Unit

120: Artificial Intelligence Identification Analysis Module

121: Artificial Intelligence Identification Analysis Program

130: Medical Knowledge Rule Module

131: Functional Medicine Anatomy Rules

132: Dynamic Medical Anatomy Rules

140: Historical Database of Aesthetic Medical Auxiliary Assessment Results

141: Results of Aesthetic Medical Auxiliary Assessments

150: Output input module

160: Electronics

220: Artificial Intelligence Identification Analysis Module

221: Deep Learning/Training Procedures for Artificial Intelligence

230: Medical Knowledge Rule Module

231: Functional Medicine Anatomy Rules

232: Dynamic Medical Anatomy Rules

240: Historical Database of Aesthetic Medical Auxiliary Assessment Results

241: Aesthetic Medical Aid Assessment Results

A: Instant facial expression evaluation results

A’: The result of facial expression evaluation

A1: Static expression evaluation results

A2: Dynamic expression evaluation results

A31-A33, A3n: Combination of sentiment indicators

AU1-AUn: facial action unit of static expression evaluation results

AU1'-AUn': facial action unit of dynamic expression evaluation results

B: Basic information

B1: Gender

B2: Age

C, C1-Cn: Combinations and Priority Orders for Assessing Treatment Site Outcomes

P: Personal facial features

P1, P1': static texture features of habitual expressions

P2,P2': static contour feature

P3, P3': skin texture characteristics

D: Types of injectable fillers

U: Dosage of injectable filler

R1, R11-R14: Plural medical rules of functional medical anatomical rules

R2, R21-R24: Plural medical rules of dynamic medical anatomy rules

S11-S14, S111-S113: Process steps of AI-assisted evaluation method

S31-S32: Process steps of artificial intelligence deep learning/training program

FIG. 1 is a conceptual diagram of a preferred implementation of the artificial intelligence-assisted evaluation system in the concept of the present invention.

FIG. 2 is a flow chart of a preferred implementation of the artificial intelligence-assisted evaluation method in the concept of the present invention.

FIG. 3 is a schematic block diagram of a preferred implementation of the facial expression evaluation module in FIG. 1 .

FIG. 4 is a flow chart of a preferred implementation of the facial expression evaluation module in FIG. 3 .

FIG. 5 is a conceptual schematic diagram of a preferred implementation of the facial motion coding unit in FIG. 3 .

FIG. 6A is a conceptual schematic diagram of a specific implementation of generating static expression evaluation results and dynamic expression evaluation results by the facial motion coding unit in FIG. 3 .

FIG. 6B is a conceptual schematic diagram of a specific implementation of quantitatively analyzing the action intensity of the facial muscle groups of each facial action unit under different emotional expressions by the emotion analysis and face recognition unit in this case.

FIG. 7A is a conceptual diagram of a preferred implementation of the artificial intelligence deep learning/training architecture of the artificial intelligence identification and analysis module in the concept of the present invention.

FIG. 7B is a conceptual diagram of a preferred implementation of the historical database of cosmetic medical assistance evaluation results applied to the artificial intelligence identification and analysis module in FIG. 7A .

FIG. 8 is a flow chart of a preferred implementation of the artificial intelligence deep learning/training procedure of the inventive concept of the present application in FIG. 7A .

FIGS. 9A to 9F are schematic diagrams illustrating a first preferred implementation of the artificial intelligence-assisted evaluation method and the system thereof according to the inventive concept of the present application.

Figure 10: It is the application of artificial intelligence assistance in the concept of the invention in this case A conceptual schematic diagram of a second preferred implementation of the evaluation method and its system.

FIG. 11 is a schematic diagram of a third preferred implementation concept of applying the artificial intelligence-assisted evaluation method and system thereof in the concept of the present invention.

The following examples are provided for detailed description, and the examples are only used as examples to illustrate, and do not limit the scope of protection of the present invention. In addition, the drawings in the embodiments omit unnecessary elements or elements that can be completed with ordinary techniques, so as to clearly show the technical characteristics of the present invention. Hereinafter, the preferred embodiments of the present invention are further described with reference to the drawings.

Please refer to FIG. 1 and FIG. 2 , which are respectively a conceptual diagram of a preferred implementation of the artificial intelligence-assisted evaluation system in the concept of the present invention, and a flow chart of a preferred implementation of the artificial intelligence-assisted evaluation method.

As shown in FIG. 1 , the artificial intelligence-assisted evaluation system 100 used in this case is based on facial expressions and applied to cosmetic medicine, including: a facial expression evaluation module 110 , an artificial intelligence identification and analysis module 120 , and a medical knowledge rule module 130 , a historical database 140 of cosmetic medical assistant evaluation results, and an output and input module 150 .

The facial expression evaluation module 110 at least includes a facial motion encoding unit 111 , an emotion analysis and facial recognition unit 112 , and a facial image capture unit 113 . The artificial intelligence identification and analysis module 120 at least includes an artificial intelligence identification and analysis program 121 . The medical knowledge rule module 130 at least includes functional medical anatomy rules 131 and dynamic medical anatomy rules 132 . The historical database 140 of aesthetic medical assistant evaluation results includes at least a plurality of aesthetic medical assistant evaluation results 1-N.

Furthermore, the input/output module 150 is used for receiving input or outputting various information, for example: receiving the basic data B1-B2 and/or personal face of the input subject The features P1-P3 are input into the artificial intelligence identification and analysis module 120; or, the combination and the preferred order C1-Cn of the results of evaluating the treatment site received from the artificial intelligence identification and analysis module 120 are output, and/or injection filling The type D of the agent and the dose U of the injected filler.

Among them, the facial expression evaluation module 110, the artificial intelligence recognition and analysis module 120 and the input and output module 150 can be assembled to form an electronic device 160, and the electronic device 160 can be a handheld smart mobile device, a personal computer (PC) or an independent A stand-alone smart device. For example, the electronic device 160 can be a tablet computer, a smart mobile device, a notebook computer, a desktop computer, a stand-alone smart device or a stand-alone smart module, wherein the smart device or the smart module can be Assembled or separated into a medical device (not shown).

Moreover, the electronic device 160 is connected to the historical database 140 and/or the medical knowledge rule module 130 in the form of wireless transmission and/or wired transmission. For example, the historical database 140 and/or the medical knowledge rule module 130 of the aesthetic medical assistant evaluation results can be stored in a cloud storage platform, and the electronic device 160 can be connected to the cloud storage platform via various regional/wide area networks (not shown in the figure). Shows).

Next, please refer to FIG. 2 , which is an AI-assisted evaluation method S10 applied in the aforementioned AI-assisted evaluation system 100 , and please refer to FIG. 1 ; that is, the aforementioned AI-assisted evaluation method S10 includes implementing the following Step: Execute step S11: provide the test subject's real-time facial expression evaluation result A, and the test subject's basic data B1-B2 and/or personal facial features P1-P3.

The real-time facial expression evaluation result A may include the static expression evaluation result A1 provided by the facial expression evaluation module 110, the dynamic expression evaluation result A2, and at least one of the static expression evaluation result A1 and the dynamic expression evaluation result A2. The result is a combination of multiple sentiment indicators A31-A3n.

In addition, the basic data B1-B2 and/or the personal facial features P1-P3 of the subject, in addition to the method provided by the input and output module 150, in other preferred embodiments, can also be evaluated by facial expressions. The module 110 directly provides the other facial features P1'-P3' of the subject; and the aforementioned methods can be selected or performed by division of labor, none of which deviates from the inventive concept of the present case.

Next, step S12 is performed: the aforementioned real-time facial expression evaluation result A, and the basic data B1-B2 of the subject and/or personal facial features P1-P3 are input into the artificial intelligence recognition analysis module 120 .

Next, step S13 is executed: the artificial intelligence identification and analysis module 120 can select at least one of the medical knowledge rule module 130 and the aesthetic medical assistant evaluation result historical database 140 to execute the artificial intelligence identification and analysis program 121, and generate and A cosmetic medical assistant evaluation result is output; wherein, the artificial intelligence identification and analysis module 120 can be connected to the medical knowledge rule module 130 and the cosmetic medical assistant evaluation result historical database 140 by wireless transmission or wired transmission.

Furthermore, the artificial intelligence identification and analysis module 120 can be based on the functional medical anatomy rules 131 and the dynamic medical anatomy rules 132 in the medical knowledge rule module 130 , and/or a plurality of the historical database 140 of the aesthetic medical assistant evaluation results. Aesthetic medical assistant evaluation results 1-N, and adaptively generate aesthetic medical assistant evaluation results corresponding to the subject; wherein, the aesthetic medical assistant evaluation results at least include the combination of the subject's evaluation treatment site results and the priority order C1- Cn.

Of course, the aforementioned cosmetic medical assistant evaluation results may further include the combination and the preferred order C1-Cn corresponding to the evaluation treatment site results, so as to generate the assistant evaluation results regarding the injection filler type D and filler dose U.

In addition, the artificial intelligence recognition and analysis module 120 can output or display the combination of the results of evaluating the treatment site and the output and input module 150 shown in FIG. Preference is given to the sequence C1-Cn and/or the type D of filler injected and the dose U of filler.

In addition, regarding another preferred implementation of step S13 in this case (not shown in the figure), the artificial intelligence recognition and analysis module 120 in this case can also respond to the real-time facial expression evaluation result A and various built-in medical knowledge rules and/or the cosmetic medical assistant evaluation result, and directly execute the aforementioned artificial intelligence identification and analysis program 121, and generate another cosmetic medical assistant evaluation result; the implementation feature of the above-mentioned another preferred method is that there is no need to additionally match the above-mentioned external medical knowledge rules. At least one of the module 130 and the historical database 140 of pre-cosmetic medical aid evaluation results.

A better practice in this case further includes performing step S14: feeding back and storing the aesthetic medical assistant evaluation result to at least one of the medical knowledge rule module 130 and the aesthetic medical assistant evaluation result historical database 140; wherein, the artificial intelligence identification The analysis module 120 can store the cosmetic medical assistant evaluation result of each subject in the historical database 140 of cosmetic medical assistant evaluation results, so as to facilitate the artificial intelligence identification and analysis module 120 according to at least one artificial intelligence deep learning/training algorithm. Execute an artificial intelligence deep learning/training program, and this part will be supplemented later, and will not be repeated here.

Of course, the aforementioned step S14 can also be chosen not to be implemented in other embodiments, and these are all equivalently modified implementation aspects of the present case.

More specifically, regarding how the facial expression evaluation module 110 in the aforementioned step S11 provides the real-time facial expression evaluation result A, the following steps may be further included. Please refer to FIG. 3 and FIG. 4 , which are respectively a conceptual diagram of a preferred implementation of the facial expression evaluation module in the concept of the present invention, and a flowchart of a preferred implementation of the facial expression evaluation module.

As shown in FIG. 3 , the facial expression evaluation module 110 of this case includes: a facial motion encoding unit 111 , an emotion analysis and facial recognition unit 112 , and a facial image capturing unit 113 ; wherein, the facial image capturing unit 113 to perform an image capture After taking the action (for example, shooting with a camera device (not shown)), a real-time facial image can be obtained and output to the facial motion encoding unit 111; after that, the facial motion encoding unit 111 will The real-time facial image and medical knowledge rule module 130 subdivides a real-time facial action presented in the image into a plurality of facial action units AU1-AUn (please refer to the related description of FIG. 5 described later).

As for the emotion analysis and face recognition unit 112, it includes a plurality of emotion index combinations A31-A3n, which may belong to a positive emotion index combination or a negative emotion index combination; wherein, the aforementioned positive emotion index combination refers to including, For example: one or a combination of happiness index, happiness index, satisfaction index, moving index, positive index and relaxation index, and the aforementioned combination of negative emotion index refers to, for example: sadness index, anger index, worry index, surprise One or a combination of indicators, indicators of fear, indicators of contempt and indicators of hatred.

Next, please refer to FIG. 4 again. Step S11 further includes the following steps. Step S111 is executed: using the facial expression evaluation module 110 and according to the medical knowledge rule module 130 to differentiate a facial muscle action into multiple faces facial action units AU1-AUn; wherein, the facial action encoding unit 111 can be used to distinguish various facial muscle actions according to the functional medical anatomy rules 131 and the dynamic medical anatomy rules 132 in the medical knowledge rule module 130, and the aforementioned A plurality of medical anatomical rules can be, for example, the linkage relationship between each muscle group and adjacent muscle groups, and/or the function of each muscle group.

For example, please refer to FIG. 5 , which is a conceptual schematic diagram of a preferred implementation of the aforementioned facial motion encoding unit 111 . As shown in FIG. 5 , the facial action coding unit 111 of this case can define the frown muscle at the brow as the facial action unit AU4 (Face Action Unit 4), or the depressor anguli oris which may cause the corner of the mouth to droop. and the mental muscle (mentails) are defined as the face action unit AU15 (Face Action Unit 15) and the face action unit AU17 (Face Action Unit 15), respectively. Action Unit 17).

Next, step S112 is continued: according to at least one of the static expression evaluation result A1 and the dynamic expression evaluation result A2, a plurality of emotion index combinations A31-A3n are formed; wherein, the facial motion encoding unit 111 is based on each facial expression The static expression evaluation result A1 and the dynamic expression evaluation result A2 are formed by the linkage between the action unit and another facial action unit in response to the change of the facial emotion or the combination of actions and the intensity of the muscle action.

Please refer to FIG. 6A together with FIG. 3 , wherein FIG. 6A is a specific implementation concept of generating the static expression evaluation result A1 and the dynamic expression evaluation result A2 by the facial motion encoding unit 111 in the facial expression evaluation module 110 of the present case Schematic diagram; wherein, the facial expression evaluation module 110 of this case first detects the changes of the subject's facial expressions in static and dynamic situations (such as the static changes in the left picture to the dynamic changes in the right picture), and then, for example, The static expression evaluation result A1 and the dynamic expression evaluation result A2 are formed according to the detected facial action unit AU1 , the facial action unit AU4 and the facial action unit AU15 .

Wherein, as shown in the left figure of FIG. 6A , the aforementioned static expression evaluation result A1 may be the static parameter values of a plurality of facial action units AU1-AUn when the subject has no emotions, or the subject can A short video is recorded in a relaxed state, which can detect and analyze whether the multiple facial muscle groups of the subject will exert force or act unconsciously, which is equivalent to the linkage and dynamics between the muscle groups.

The aforementioned dynamic expression evaluation result A2, as shown on the right side of FIG. 6A, is that the subject presents different facial expressions according to different emotions, for example, it may be: sadness; of course, the aforementioned dynamic expression evaluation result A2 You can also include more other angry expressions, laughing expressions, etc. different emotional expressions.

Next, please refer to FIG. 6B with FIG. 3 , wherein FIG. 6B is The emotion analysis and face recognition unit 112 in this case is used to further quantitatively analyze the action intensity of the facial muscle group of each facial action unit under different emotional expressions, so as to provide a schematic diagram of a specific implementation concept of more accurate dynamic parameter values. That is, the emotion analysis and face recognition unit 112 of this case can form the aforementioned multiple emotion index combinations A31-A3n according to at least one of the static expression evaluation result A1 and the dynamic expression evaluation result A2.

The facial expression evaluation module 110 in this case can, for example, divide facial expressions into 7 categories of definitions, which, in addition to neutral expressions, also include: happy (Happy), sad (Sad), angry ( Angry), Surprise (Surprise), Fear (Fear/Scared), Disgusted (Disgusted) and other expression definitions, which is equivalent to the combination A31-A3n of various emotional indicators mentioned above in this case.

Wherein, the facial action encoding unit 111 in this case defines the features of sad emotional expressions as Eyebrow down, Frown line, Clock time eyebrow 8:20, Reverse U, Brow head up, Dark eyes, Lower lip protruding, Lip tight, Lip corners droop, Chin bulges, etc. The encoding of the facial action unit, and the features of the above-mentioned sad emotional expression, can be used as the basis for the emotion analysis and face recognition unit 112 to determine the sadness index and distinguish the sadness index from other emotion indexes (not shown).

That is to say, in this case, multiple facial action units AU1-AUn can be used, and by identifying the emotional expression characteristics related to emotions comprehensively expressed by various facial muscle groups, and comprehensively analyzing the facial expressions to be presented or conveyed real emotions. In this way, the emotion analysis and face recognition unit 112 can objectively identify and analyze the subtle changes in facial expressions by using the micro-expressions between the static expression evaluation result A1 and the dynamic expression evaluation result A2, and thus be more accurate. It is quantified by evaluating the actuation among a plurality of facial action units AU1-AUn.

Of course, in different preferred embodiments, the emotion analysis and face recognition unit can be assembled and combined or installed and loaded into at least one of the facial expression evaluation module and the artificial intelligence recognition analysis module, so that it can be based on At least one of the static expression evaluation result and the dynamic expression evaluation result forms a combination of the above-mentioned multiple emotion indicators.

Finally, step S113 is executed: the real-time facial expression evaluation result A is formed according to the ratio result of each combination of the emotion indicators; wherein, the real-time facial expression evaluation result A includes the ratio results of multiple emotion indicator combinations A31-A3n, such as : The angry index is 14.1%, the sadness index is 35.2%...etc.

Continuing to describe how the artificial intelligence identification and analysis module 120 for identifying and analyzing the real-time facial expression evaluation result A performs artificial intelligence deep learning/training. Please refer to FIG. 7A and FIG. 7B , which are respectively a conceptual diagram of a preferred implementation of the AI deep learning/training framework of the AI recognition analysis module, and the historical data of the aesthetic medical aid evaluation results applied to the AI recognition analysis module Conceptual diagram of the preferred implementation of the library.

As shown in FIG. 7A and FIG. 7B , the artificial intelligence deep learning/training framework of the artificial intelligence identification and analysis module includes: an artificial intelligence identification and analysis module 220 , a medical knowledge rule module 230 , and a history database 240 of cosmetic medical aid evaluation results; The artificial intelligence identification and analysis module 220 includes an artificial intelligence deep learning/training program 221 , the medical knowledge rule module 230 includes functional medical anatomy rules 231 and dynamic medical anatomy rules 232 , and a history database 240 of cosmetic medical assistant evaluation results. Multiple cosmetic medical aid assessment results 241 are included.

Among them, each cosmetic medical assistant evaluation result at least includes the name of the subject, basic information B, facial expression evaluation result A' at the time of the test, personal facial feature P, functional medical anatomy rules 231 and dynamic medical The combination of multiple medical rules R1, R2 in anatomical rules 232, evaluating the results of the treatment site and The preferred combination of C1, C2 in the preferred order and the type D of the injectable filler and the dose U of the filler and so on.

In practical application, the basic data B may include gender B1 and age B2; the facial expression evaluation result A' includes at least: static expression evaluation result A1, dynamic expression evaluation result A2, and multiple emotional index combinations A31-A33; The expression evaluation result A1 can be the static parameter values of the multiple facial action units AU1-AUn when the subject has no emotions; the dynamic expression evaluation result A2 can be the multiple facial movements generated by the subject according to different emotions For each dynamic parameter value of the units AU1'-AUn', the emotion index combination A31-A33 may include, for example, the fear index A31, the anger index A32 and the contempt index A33 of the negative emotion index combination, or may include the happiness index A31 of the positive emotion index combination , moving index A32 and meeting index A33.

Furthermore, the personal facial features P may include static texture features P1 of habitual expressions, static contour features P2 or skin texture features P3; wherein, the personal facial features P1-P3 can be obtained by the facial expression evaluation module 110 and the input and output modules. provided by at least one of the groups 150 .

As for the medical rule R1 of the functional medical anatomy rule 231, for example, it includes the stretching degree rules of each facial muscle group in response to different emotional expressions, and the tension degree rules R11-R14; and the medical rule R2 of the dynamic medical anatomy rule 232, such as Including the linkage rules and contraction rules R21-R24 between various facial muscle groups in response to different emotional expressions.

Finally, the preferred combination C1, C2 can be, for example, one of the multiple facial action units AU1-AUn of the subject to be treated or a combination thereof; the type D of injectable fillers can include: hydrogel type W , botulinum toxin agent class X, hyaluronic acid agent class Y and collagen agent class Z. Among them, hydrogel agent W, hyaluronic acid agent Y and collagen agent Z, in addition to reducing the static state of human facial expression In addition to reducing the negative emotion index combination (sadness index, anger index, etc.), it can also increase the positive emotion index combination (happy index, satisfaction index, etc.).

Of course, in this embodiment, the aforementioned cosmetic medical assistant evaluation results 1-N can be adjusted according to the needs of the actual cosmetic medical treatment target, which should not be limited to this example. In addition, the content of the aforesaid cosmetic medical assistant evaluation results 1-N can be modified or designed by those skilled in the art, and can be adjusted and designed according to the actual needs of the subject's cosmetic medical treatment.

Continuing to describe how the AI recognition and analysis module 220 executes the AI deep learning/training program 221 according to the above-mentioned AI deep learning/training framework. Please refer to FIG. 8 , which is a flow chart of a preferred implementation of the aforementioned artificial intelligence deep learning/training procedure.

As shown in FIG. 8 , the process steps of the artificial intelligence deep learning/training program 221 include: performing step S31 : providing the cosmetic medical assistant evaluation results 241 ; wherein each of the cosmetic medical assistant evaluation results at least includes: historical subjects Relevant basic data B, facial expression evaluation results A', personal facial features P, functional medical anatomy rules 231 and dynamic medical anatomy rules 232 in the medical knowledge rule module 230, and the combination of the results of evaluating treatment parts and The preferred order and the type D of the injection filler and the dosage U of the filler; wherein, the content of the cosmetic medical assistant evaluation results 241 are those shown in the aforementioned FIG. 7B , of course, the static expression evaluation result A1 and the dynamic expression evaluation result A2 can be Choose whether to record it or not according to the actual use requirements.

Afterwards, step S32 is executed: inputting the cosmetic medical assistant evaluation results 241 to the artificial intelligence identification and analysis module 220 for it to execute the artificial intelligence deep learning/training program 221; wherein the artificial intelligence identification and analysis module 220 is based on at least one The AI deep learning/training algorithm is executed while the AI deep learning/training program 221 is executed. The artificial intelligence deep learning/training algorithm is Machine learning algorithm, artificial neural network road algorithm, fuzzy logic algorithm, deep learning algorithm or a combination thereof, in this case, at least one of artificial neural network road algorithm and deep learning algorithm is preferred.

Furthermore, a number of preferred implementation concepts will be described later, how to use the artificial intelligence-assisted evaluation method and system in this case to perform cosmetic medical behaviors. Among them, the main treatment target requirements of the following preferred embodiments are to improve the negative emotion index combination of the facial expressions of the subjects 1-3, and by reducing or improving the negative emotion index combination caused by the negative expression emotion, This enhances personal charisma and interpersonal relationships.

For example, the main therapeutic purpose of these preferred embodiments is to reduce or improve involuntary frowning or drooping facial expressions, so as to reduce the feeling of being angry or serious, and even reduce negative micro-expressions, and also In other preferred embodiments, the part of the combination of positive emotion indicators can be further strengthened, so as to achieve a higher-quality, more accurate and personalized aesthetic beauty medical effect.

Please refer to FIG. 9A to FIG. 9F , which are schematic diagrams of a first preferred implementation concept of applying the artificial intelligence-assisted evaluation method and system thereof in the inventive concept of the present application.

Taking the contents of FIGS. 9A to 9F in combination with FIGS. 1 to 4 and FIG. 7B as an example, the facial expression evaluation module 110 detects a plurality of facial action units AU1-AUn of the subject 1 , and according to the change between a detection result of each facial action unit AU1-AUn and another detection result of another facial action unit AU1-AUn, a real-time facial expression evaluation result A is provided; wherein, The real-time facial expression evaluation result A is formed according to the ratio result of each emotion index combination generated according to the static expression evaluation result A1 and the dynamic expression evaluation result A2.

As shown in Figure 9A, the subjects may frown unconsciously, or the corners of the mouth may sag due to aging. These facial action units AU1-AUn The resulting micro-expressions are recorded, and a static expression evaluation result A1 is formed.

In addition, as shown in FIG. 9B , the dynamic expression evaluation result A2 is that the subject 1 presents different facial expressions according to different emotions, such as angry expressions, laughing expressions, etc.

Next, the emotion analysis and face recognition unit 112 further quantitatively analyzes the action intensity of the facial muscles of each facial action unit under different emotional expressions (including the aforementioned static expression evaluation result A1 and dynamic expression evaluation result A2), and More accurate dynamic parameter values are provided to be used as the emotional index combination A31-A33, the combination and priority order C1-C2 for evaluating the results of the treatment site, and the type D of the injected filler and the dose U of the filler. The treatment reference.

Of course, in other preferred embodiments, the emotion analysis and face recognition unit may be a process step, which is built-in and installed in the AI recognition analysis module, or is a part of the AI recognition analysis program program (not shown).

Furthermore, as shown in FIG. 9C and FIG. 9D , the multiple emotion index combinations A31-A33 of the subject 1 are respectively 35.2% for the sadness index, 14.1% for the anger index, and 17.7% for the fear index. Relevant information of each facial action unit AU1-AUn corresponding to the index combination A31-A33.

On the other hand, the facial expression evaluation module 110 can also be equipped with a face three-dimensional (3D) simulation unit and a skin texture detection unit to further provide the static texture feature P1 and the static contour feature P2 of the habitual expression of the personal facial feature P. Or skin texture feature P3, etc. (not shown in the figure).

Then, as shown in FIG. 9E , input the real-time facial expression evaluation result A into the artificial intelligence recognition and analysis module 120, or the artificial intelligence recognition and analysis module 120 actively receives the real-time facial expression evaluation result A, and selects whether to match the medical evaluation result A. At least one of the treatment knowledge rule module 130 and the aesthetic medical assistant evaluation result historical database 140 is used to execute the artificial intelligence identification and analysis program 121 . After that, generate and output the cosmetic medical assistant evaluation result of the subject 1, wherein the cosmetic medical assistant evaluation result at least includes the combination of the evaluation results of the treatment site and the preferred order C1-C2, and the type D of the injected filler and the dose U of the filler .

For this example, Botulinum Toxin 8 sU was administered to the muscle groups related to the facial action unit AU1 (medial frontalis, Inner Frontalis), and to the muscle groups related to the facial action unit AU15 (lower frontalis). The angle of the mouth muscle (depressor anguli oris) is a muscle group (like muscle, mentals) related to the facial action unit A17, and the results of the cosmetic medical assistant evaluation of Botulinum Toxin DAO 4 sU and Mentalis 4 sU are recommended.

In this way, please refer to FIG. 9F , compare the facial expression evaluation results A′ of the subject 1 before, one week and three weeks after the treatment, from which it can be found that the face of the subject 1 The sadness index decreased from 35.2% to 0% directly after one week of treatment; in addition, the change of the angry index on the face of subject 1 decreased from 14.1% before treatment to one week after treatment. It was 7.8%, and even three weeks after the botox treatment, the facial vitality index would completely drop to 0%.

On the other hand, the evaluation results of aesthetic medical assistance for sadness indicators suggest that through the implementation and operation of this case, another treatment reference guidance path can also be given: for example, when the sadness index accounts for 10% of the total emotional index combination (total expression) When the above, and the facial motion unit AU1, facial motion unit AU4, and facial motion unit AU15 in the facial motion coding unit are all enhanced (that is, the proportion is increased), it is recommended to inject at the corresponding muscle group position Botulinum Toxin A type; Of course, in this case, by using more case data (beauty medical assistant evaluation results) and performing multiple artificial intelligence deep learning/training procedures, it is obvious that a better treatment proposal can be proposed. With the above-mentioned cosmetic medical aids assessment Results are limited.

Furthermore, please refer to FIG. 10 , which is a conceptual schematic diagram of a second preferred implementation of the artificial intelligence-assisted evaluation method and system thereof according to the concept of the present invention.

As shown in FIG. 10 , the facial expression evaluation module 110 detects a plurality of facial action units AU1-AUn of the subject 2, so as to provide a real-time facial expression evaluation result A, and thus the subject is known. Among the multiple emotional index combinations of 2, the neutral index is the highest at 26.3%, and the sadness index is the second highest at 13.9%. At the same time, in the auxiliary evaluation results of beauty medical treatment, the order of the recommended treatment sites is as follows: The facial action unit AU1 of the sadness indicator is classified as Inner Brow Raiser.

Based on this, the aforementioned real-time facial expression evaluation result A is combined with the medical knowledge rule module 130 and the aesthetic medical auxiliary evaluation result historical database 140, and it can be known that the functional medical anatomy rules and the dynamic medical anatomy rules can guide The positions of the muscle groups that are highly correlated and linked with the aforementioned facial action unit AU1 can be obtained, and then the aesthetic medical assistant evaluation results for personalised cosmetic medical treatment can be further provided, for example, for the muscles related to the facial action unit AU1 Group (medial frontalis, Inner Frontalis), reference recommendations for treatment with botulinum toxin Toxin 8 sU.

In this way, comparing the facial expression evaluation results A' of the facial expression of the subject 2 before, one week and three weeks after the treatment, it can be found that the sadness index of the subject 2's face is from 13.9%. After up to one week of treatment, it dropped to 8.4% directly, and even three weeks after Botox treatment, the sadness index on the face could be completely reduced to 0%. In short, through the artificial intelligence-assisted evaluation method and its system applied to beauty medical treatment in this case, the cosmetic medical treatment effect that can be obtained is indeed very significant.

After that, please refer to FIG. 11 , which is a conceptual schematic diagram of a third preferred implementation of the artificial intelligence-assisted evaluation method and system thereof in the inventive concept of the present application.

As shown in FIG. 11 , the facial expression evaluation module 110 detects a plurality of facial action units AU1-AUn of the subject 3, so as to provide a real-time facial expression evaluation result A. From this, it can be known that whether Which gender or age the human angry expression is is roughly related to the aforementioned muscle groups of the facial action unit AU15 and the facial action unit AU17; and compare the facial expressions of the subject 3 before, one week and three days after the treatment. The facial expression evaluation result A' after one week, from which it can be found that the anger index of the subject 3's face has been significantly improved after three weeks of treatment.

That is to say, it is suggested that another treatment reference guide path can be given through the implementation and operation of this case, for example, when the anger index is more than 10%, and the facial action coding unit is When the facial action unit AU15 and the facial action unit AU17 are both enhanced (that is, the proportion increases), it is recommended to inject Botulinum Toxin A type into the corresponding muscle groups (depressor depressor muscle and mental muscle).

However, in view of the above-mentioned preferred embodiment, and comparing with the current conventional cosmetic medical treatment recommendations based only on the personal judgment of physicians, it can be seen that the conventional method is indeed easily limited by the influence of physicians’ personal experience and stereotypes. , so that it cannot be considered comprehensively and objectively, and there is a major defect of missing the differences in each person's microexpressions.

In detail, the target of treatment is to reduce the index of anger. Physicians often use Botox to reduce the movement of the frown muscle, but ignore that the muscle movement of each person's angry expression is actually slightly different. A few people may have the corners of the mouth down at the same time, or the muscles of the chin. It will contract and rise, and some people have some movements on the inner side of the eyebrow raising muscle, and the It is possible that only part of the movement is visible to the naked eye, so it is easy to cause the muscle movement to be too subtle and difficult to detect, resulting in blind spots and misunderstandings in treatment, resulting in adverse medical results and unnecessary medical disputes.

For example, if the subject 1 in FIG. 9A to FIG. 9F is changed to a cosmetic medical treatment recommendation based on the personal judgment of a conventional doctor, the main treatment parts will be concentrated on the facial action unit AU15 and the facial action. In the unit AU17, the doctor omits the facial action unit AU1 of the anger index, and the effect of cosmetic medical treatment is not good (not shown in the figure).

Next, in the subject 2 in FIG. 10, the cosmetic medical treatment recommendation based on the personal judgment of a conventional doctor is that the main treatment part is usually the facial action unit AU2, and the judgment is based on the orbicularis oculi of the subject 2. Drooping eyes, which cause sadness indicators. However, after performing the cosmetic medical behavior, comparing the facial expression evaluation results A' of the facial expression of the subject 2 before, one week and three weeks after the treatment, the sadness indicators on the face of the subject 2 can be found. , it dropped from 6.8% to 5.1% after one week of treatment, but the sadness index returned to 6.7% three weeks after botulinum toxin treatment. As a result, the effect of cosmetic medical treatment is not good, so that the sadness index cannot be effectively improved (not shown in the figure).

Finally, for subject 3 in Figure 11, if the cosmetic medical treatment recommendation based on the personal judgment of a conventional physician is changed, the main treatment site is usually the injection of botulinum toxin abobotulinumtoxin A at the facial action unit AU17 for a total of 4 doses (Units). However, after performing the cosmetic medical behavior, compare the facial expression evaluation results A' of the facial expression of the subject 3 before, one week and three weeks after the treatment, from which the anger index of the subject 3's face can be found , it dropped from 10.9% to 5.9% after one week of treatment, but the anger index rose to 13.9% three weeks after botulinum toxin treatment. The reason is that the doctor not only missed to judge the treatment site of subject 3 (Facial Action Unit AU15), and injected with Botox The dose of abobotulinumtoxin A in the facial action unit AU17 was also insufficient, which not only failed to improve the subject's facial expression, but also had the opposite effect of increasing the anger index (not shown in the figure).

Therefore, compared with the artificial intelligence-assisted evaluation method and system used in beauty medicine in this case, the facial expression evaluation module 110 provides the real-time facial expression evaluation result A of the subject, and then recognizes it through artificial intelligence. The analysis module selects and matches the plural medical rules in the medical knowledge rule module and the historical database of the aesthetic medical assistant evaluation results to execute the artificial intelligence identification and analysis program, thereby generating and outputting the aesthetic medical assistant evaluation results, of which the aesthetic medical assistant evaluation results are at least Include the subject's assessment of the combination and preferred order of treatment site results and/or the type and dose of filler injected. In this way, this case can not only accurately analyze and evaluate the correct and complete treatment site (facial action unit AU1-AUn), but also accurately provide the type and dosage of injectable fillers, so as to achieve personalized aesthetic beauty treatment. treatment effect.

In addition, this case can also carry out cosmetic medical behaviors that strengthen the combination of positive emotional indicators, or cosmetic treatment suggestions for preventive improvement of the aging treatment area of the face. For example, after the facial muscles relax, the corners of the mouth will sag, so The faces that may generate angry indicators are equal.

On the other hand, the method and system of this case can be applied to a variety of cosmetic medical or aesthetic fields, and can also be used as a basis for judging the treatment effect before and after cosmetic medical surgery. blind spots and mistakes.

The above descriptions are only the preferred embodiments of the present invention, and are not intended to limit the scope of the patent application of the present invention. Therefore, all other equivalent changes or modifications made without departing from the spirit disclosed in the present invention shall be included in this case. within the scope of the patent application.

S10, S11-S14: Process steps

Claims (30)

An artificial intelligence-assisted evaluation method used in cosmetic medicine, which at least includes the following steps: providing a real-time facial expression evaluation result of a subject; inputting the real-time facial expression evaluation result into an artificial intelligence recognition analysis module; and select and match at least one of a medical knowledge rule module and a history database of cosmetic medical aid evaluation results to execute an artificial intelligence identification and analysis program, and generate and output a cosmetic medical aid evaluation result; wherein the real-time person The facial expression evaluation result includes: a static expression evaluation result, or the static expression evaluation result and a dynamic expression evaluation result, and the aforementioned providing the real-time facial expression evaluation result further includes the following steps: according to the medical knowledge rule module, make a The face is divided into plural facial action units; according to at least one of the static expression evaluation result and the dynamic expression evaluation result, a plurality of emotion index combinations are formed; and the real-time emotion index combination is formed according to the proportional result of each emotion index combination. Facial expression evaluation results. The artificial intelligence-assisted evaluation method described in item 1 of the scope of the application further comprises the following steps: feeding back and storing the aesthetic medical assistant evaluation result to the medical knowledge rule module and the aesthetic medical assistant evaluation result historical database at least one. The artificial intelligence-assisted evaluation method as described in item 1 of the scope of the application, wherein the output of the auxiliary evaluation result of cosmetic medical treatment at least includes: the combination and the priority order of the evaluation results of one of the subjects to evaluate the treatment part, or the evaluation of the treatment part Combinations and preferred order of results and an injectable filler type and dose. The artificial intelligence-assisted evaluation method as described in item 1 of the patent application scope, wherein the medical knowledge rule module further includes a functional medical anatomy rule and a dynamic medical anatomy rule. The artificial intelligence-assisted evaluation method as described in item 1 of the patent application scope, wherein the emotion index combinations are at least one of a positive emotion index combination and a negative emotion index combination. The artificial intelligence-assisted evaluation method as described in item 5 of the patent application scope, wherein the combination of negative emotion indicators is a sadness indicator, an anger indicator, a worry indicator, a surprise indicator, a fear indicator, a contempt indicator, and a contempt indicator. Hate one or a combination of indicators. The artificial intelligence-assisted evaluation method as described in item 5 of the scope of the patent application, wherein the combination of positive emotion indicators is one of a happiness index, a happiness index, a satisfaction index, a moving index, a positive index and a relaxation index one or a combination thereof. The AI-assisted assessment method as described in item 1 of the scope of the application, wherein the historical database of cosmetic medical aid assessment results includes a plurality of cosmetic medical aid assessment results, which are used for the AI identification and analysis module to be based on at least one AI A deep learning/training algorithm executes an artificial intelligence deep learning/training program. The artificial intelligence-assisted evaluation method described in item 8 of the patent application scope, wherein the artificial intelligence deep learning/training program includes the following steps: providing the aesthetic medical assistant evaluation results; wherein, each of the aesthetic medical assistant evaluation results at least Including: a basic information related to a historical subject, a person's facial expression evaluation result, a person's facial features, a functional medical anatomy rule in the medical knowledge rule module, a dynamic medical anatomy rule, an evaluation treatment part The combination and the preferred order of the results and the type and dosage of an injected filler; and inputting the aesthetic medical assistant evaluation results into the artificial intelligence identification and analysis module. The artificial intelligence-assisted evaluation method as described in item 9 of the patent application scope, wherein the personal facial feature includes a static texture feature of a habitual expression, a static contour feature or a skin texture feature. The AI-assisted evaluation method according to claim 10, wherein the at least one AI deep learning/training algorithm is at least one of an artificial neural network road algorithm and a deep learning algorithm. An electronic device using the artificial intelligence-assisted evaluation method described in item 1 of the scope of application, the electronic device at least includes: a facial expression evaluation module for providing the real-time facial expression evaluation result; the artificial intelligence recognition An analysis module, including the artificial intelligence identification analysis program, the artificial intelligence identification analysis program receives the real-time facial expression evaluation result, and generates the beauty medical aid evaluation result; and an input and output module for outputting the beauty medical aid evaluation result; Wherein, the artificial intelligence recognition and analysis module receives at least one human facial feature from at least one of the facial expression evaluation module and the input and output module. The electronic device described in item 12 of the scope of the application is connected to the historical database of the aesthetic medical assistant evaluation results and the medical knowledge rule module by at least one of a wireless transmission method and a wired transmission method. at least one. The electronic device as described in item 12 of the scope of the patent application is a handheld smart mobile device, a personal computer or an independently operated smart device. An artificial intelligence-assisted evaluation system for beauty medicine, which at least includes: a facial expression evaluation module for providing a real-time facial expression evaluation result of a subject; and an artificial intelligence identification and analysis module, which is connected to the facial expression Evaluation module; wherein, the artificial intelligence recognition and analysis module receives the real-time facial expression evaluation result, and according to at least one of a connected medical knowledge rule module and a history database of aesthetic medical assistant evaluation results, to Execute an artificial intelligence recognition analysis program, and adaptively generate and output a beauty medical assistant evaluation result; and, wherein, the real-time facial expression evaluation result includes a static expression evaluation result, or the static expression evaluation result and a dynamic expression evaluation As a result, at least one of the facial expression evaluation module and the artificial intelligence recognition analysis module further includes: an emotion analysis and face recognition unit, based on at least one of the static expression evaluation result and the dynamic expression evaluation result One, to form a combination of plural sentiment indicators; wherein, according to these sentiment indicators The proportional results of each of them are combined to form the real-time facial expression evaluation result. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 15 of the patent application scope, wherein the artificial intelligence identification and analysis module feeds back and stores the cosmetic auxiliary evaluation result to the medical knowledge rule module and the cosmetic medical auxiliary evaluation result At least one of the historical databases. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 15 of the scope of the patent application, wherein the auxiliary evaluation result for cosmetic medical treatment at least includes the combination and priority of the results of one of the subjects to evaluate the treatment part, or the evaluation of the treatment part Combinations and preferred order of results and an injectable filler type and dosage. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 15 of the patent application scope, wherein the medical knowledge rule module further includes a functional medical anatomy rule and a dynamic medical anatomy rule. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 15 of the patent application scope, wherein the facial expression evaluation module includes: a facial image capture unit for performing an image capture action to obtain a a real-time facial image; and a facial motion coding unit, according to the real-time facial image and the medical knowledge rule module, to distinguish a real-time facial motion presented in the image into a plurality of facial motion units; wherein, according to The change between one detection result of each of the facial motion units and another detection result of at least one other of the facial motion units forms the static The dynamic expression evaluation results and the dynamic expression evaluation results. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 15 of the patent application scope, wherein the combination of emotion indicators is at least one of a combination of positive emotion indicators and a combination of negative emotion indicators. According to the artificial intelligence-assisted evaluation system for beauty medicine as described in item 15 of the patent application scope, the combination of negative emotion indicators is a sadness indicator, an anger indicator, a worry indicator, a surprise indicator, a fear indicator, and a contempt indicator One or a combination of an indicator and a nuisance indicator. The artificial intelligence-assisted evaluation system for beauty medicine as described in item 15 of the patent application scope, wherein the combination of positive emotion indicators is a happiness indicator, a happiness indicator, a satisfaction indicator, a moving indicator, a positive indicator and a relaxation indicator One or a combination of indicators. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 15 of the scope of the patent application, wherein the historical database of cosmetic medical auxiliary evaluation results includes a plurality of cosmetic medical auxiliary evaluation results, which are used for the artificial intelligence identification and analysis module to be based on at least An AI deep learning/training algorithm executes an AI deep learning/training program. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 23 of the patent application scope, wherein each auxiliary evaluation result for cosmetic medical treatment at least includes: a basic information related to a historical subject, a facial expression evaluation result of a person, a person One of facial features, a functional medical anatomy rule and a dynamic medical anatomy rule in the medical knowledge rule module, a combination and priority order of evaluating treatment site results, and a type and dose of injected filler, or and its combinations. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 24 of the patent application scope, wherein the personal facial feature includes a static texture feature of a habitual expression, a static contour feature or a skin texture feature. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 24 of the patent application scope, wherein the personal facial features are provided from at least one of the facial expression evaluation module and an input/output module. to the artificial intelligence identification and analysis module. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 23 of the patent application scope, wherein the artificial intelligence deep learning/training program is to input the cosmetic medical auxiliary evaluation results into the artificial intelligence identification and analysis module. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 23 of the patent application scope, wherein the at least one artificial intelligence deep learning/training algorithm is at least one of an artificial neural network road algorithm and a deep learning algorithm By. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 15 of the patent application scope, which is assembled to form an electronic device by the facial expression evaluation module, an input and output module, and the artificial intelligence recognition and analysis module; Wherein, the electronic device is a handheld smart mobile device, a personal computer or an independently operated smart device. The artificial intelligence-assisted evaluation system for cosmetic medicine as described in item 29 of the patent application scope, wherein the electronic device is connected to the historical data of the auxiliary evaluation results for cosmetic medicine by at least one of a wireless transmission method and a wired transmission method at least one of the library and the medical knowledge rule module.

Images