Facial feature recognition and remodeling optimization method for plastic surgery

Abstract

The invention relates to the technical field of facial expression image recognition, in particular to a facial feature recognition and remodeling optimization method for plastic surgery. The method comprises the steps of obtaining key points in an expression image, and performing region division; according to the key point distribution, the edge distribution and the region distribution in each region, obtaining information rich indexes, and grading the regions; obtaining a basic stability factor according to distance features and regional differences between key points in the natural expression image and corresponding points of other images; obtaining a structural similarity degree in combination with the regional level difference and the gray feature difference; combining to obtain a final stable factor; obtaining an optimized face model of the patient according to the final stable factor; and performing plastic optimization on the face of the patient according to the optimized face modeling. According to the method, the linkage relation of the face areas in the expression change process is considered, so that more accurate face 3D modeling is obtained, and the face of the patient can obtain a better adjusting effect.

Description

Technical Field

[0001] The present invention relates to the technical field of facial expression image recognition, and in particular to a facial feature recognition and reshaping optimization method for plastic surgery. Background Art

[0002] Plastic surgery refers to the process of repairing, reconstructing, or improving the human body through surgical intervention. It is typically used in medical fields such as cosmetic surgery and wound repair. Facial plastic surgery is a common type of surgery, often performed through surgical or non-surgical methods to modify and reshape the face and improve its overall appearance. Before plastic surgery, computer vision technology is often used to extract facial features and reconstruct a 3D facial model. Finally, a personalized facial plastic surgery plan is developed based on the 3D model, and the post-operative results are simulated to improve the success rate of the surgery and patient satisfaction.

[0003] When creating a 3D model of a patient's face, a single or multiple facial images are typically used, depending on the specific technology and requirements, to extract facial features, resulting in a relatively accurate 3D model. However, in facial plastic surgery scenarios, this 3D model generated from static images cannot capture the dynamic changes in facial features, nor does it consider the interactions between various facial tissues. This can lead to unpredictable postoperative consequences, such as unnatural facial muscle movements, a stiff and unrealistic appearance, and ultimately poor surgical results. Summary of the Invention

[0004] In order to solve the technical problem that during facial plastic surgery, the 3D model generated by relying on static facial expression images cannot capture the dynamic changes of facial features, nor does it take into account the linkage relationship between various facial tissues, thereby resulting in unnatural appearance of the patient's face, the purpose of the present invention is to provide a facial feature recognition and reshaping optimization method for plastic surgery, and the technical scheme adopted is as follows: a facial feature recognition and reshaping optimization method for plastic surgery, the method comprising: obtaining expression images of the patient's facial state, the expression images are divided into natural expression images in natural facial state and special expression images of all other facial states; performing key point detection on all expression images to obtain key points in each expression image; using the key points to perform region division on all expression images to obtain all facial regions in each expression image; optionally selecting a facial region in an expression image as a target region; and selecting a facial region according to the distribution of key points, edge distribution and its distribution within the target region. The invention discloses a method for determining the distribution characteristics of adjacent facial areas to obtain an information richness index of the target area; the facial areas are graded according to the information richness index to obtain the regional level of each facial area; a key point in the natural expression image is optionally used as a reference point; the corresponding point of the reference point in each special expression image is obtained according to the similarity characteristics between the special expression image and the natural expression image; the basic stability factor of the reference point in each special expression image is obtained according to the distance characteristics between the reference point and the corresponding point and the facial area difference; the structural similarity of the facial area where the reference point is located in each special expression image is obtained according to the distance characteristics, regional level difference and grayscale feature difference within the region between the facial area where the reference point is located and the facial area of each special expression image; the final stability factor of the reference point is obtained according to the basic stability factor and the structural similarity; the optimized facial modeling of the patient is obtained according to the final stability factor; and the patient's face is plastically optimized according to the optimized facial modeling.

[0005] Furthermore, the method for obtaining the information richness index includes: obtaining the information richness index according to an information richness index calculation formula, and the information richness index calculation formula is as follows: Where, Information richness indicator representing the target area; Indicates the number of key points in the target area; Indicates the area of the target region; Indicates the number of edge pixels in the target area; Indicates the The gradient value of edge pixels; Indicates the number of facial regions adjacent to the target region.

[0006] Furthermore, the method for obtaining the regional level of each facial area includes: setting a preset first number of preset threshold ranges; sorting all the preset threshold ranges from small to large, and marking a serial number for each preset threshold range as the regional level to which each preset threshold range belongs; dividing each facial area in the expression image into the preset threshold range to which it belongs according to the information richness index, and obtaining the regional level of each facial area.

[0007] Furthermore, the method for obtaining the corresponding point of the reference point in each special expression image includes: selecting any special expression image as a comparison image; marking the corresponding position of the reference point in the comparison image, and taking the key point closest to the corresponding position of the reference point in the comparison image as the corresponding point of the reference point.

[0008] Furthermore, the method for obtaining the basic stability factor includes: obtaining the basic stability factor according to a basic stability factor calculation formula, and the basic stability factor calculation formula is as follows: Where, Indicates the basic stability factor of the reference point in each special expression image; Indicates the distance between the corresponding position of the reference point and the position of the comparison point in each special expression image; Represents the area of overlap between the facial region to which the reference point belongs and the facial region to which the corresponding point belongs; The area of the largest facial region between the reference point and the corresponding point; Represents an exponential function with a natural constant as its base.

[0009] Furthermore, the method for obtaining the degree of structural similarity includes: obtaining a corresponding area of the area to which the reference point belongs in the comparison image; selecting a preset second number of facial areas closest to the area to which the reference point belongs as neighboring areas of the facial area where the reference point is located; selecting a preset second number of facial areas closest to the area to which the corresponding point belongs as neighboring areas of the facial area where the corresponding point is located; and obtaining the degree of structural similarity according to a structural similarity calculation formula, wherein the structural similarity calculation formula is as follows: Where, Indicates the structural similarity of the facial region where the reference point is located in each special expression image; Indicates the number of neighboring regions of the facial region where the reference point is located; Indicates the facial region where the reference point is located. The center pixel of the neighboring area and the corresponding point in the facial area The distance between the center pixels of adjacent areas; Indicates the facial region where the reference point is located. The neighboring area and the facial area where the corresponding point is located Differences in regional levels between adjacent regions; Indicates the grayscale mean difference between the facial area where the reference point is located and the facial area where the corresponding point is located; represents the absolute value function; Represents an exponential function with a natural constant as its base.

[0010] Furthermore, the method for obtaining the final stability factor includes: obtaining the final stability factor according to a final stability factor calculation formula, and the final stability factor calculation formula is as follows: Where, represents the final stability factor of the reference point; Indicates the number of special expression images; Indicates the reference point Basic stability factor of special expression images; Indicates the facial area where the reference point is located in The degree of structural similarity in the special expression images; Represents the normalization function.

[0011] A facial feature recognition and reshaping optimization system for plastic surgery, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of a facial feature recognition and reshaping optimization method for plastic surgery as claimed in any one of claims 1 to 7.

[0012] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned facial feature recognition and reshaping optimization method for plastic surgery.

[0013] A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned method for facial feature recognition and reshaping optimization for plastic surgery are implemented.

[0014] The present invention has the following beneficial effects: the present invention obtains facial expression images of a patient's facial state to facilitate 3D modeling of the face; in order to better capture the dynamic features of the face and ensure the accuracy and authenticity of the plastic surgery, it is necessary to identify and classify key areas of the patient's facial image to obtain more accurate facial features and classify the facial areas; the muscle movements of the face under different expressions vary from person to person, and each person's facial expressions and dynamic changes have subtle differences. In order to make the constructed facial model more accurate and personalized, it is necessary to extract change characteristics based on the change patterns of each facial area under different expressions. Therefore, the facial area where the key points are located is analyzed and a corresponding stability factor is calculated to serve as a parameter basis for constructing the facial model; because the facial change amplitude of each unnatural expression is different and the impact on the key points is also different, when calculating the final stability factor, all basic stability factors cannot be simply combined. The change amplitude of each expression must also be considered. Therefore, the structural similarity of the facial area where the reference point is located in each special expression image is analyzed to obtain a final stability factor of the reference point; an optimized facial model of the patient is obtained based on the final stability factor; and plastic surgery optimization of the patient's face is performed based on the optimized facial model. The present invention takes into account the linkage relationship between facial areas during expression changes, thereby obtaining more accurate facial 3D modeling, so that the patient's face can achieve better adjustment effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 This is a flow chart of a facial feature recognition and reshaping optimization method for plastic surgery provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0017] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of a facial feature recognition and reshaping optimization method for plastic surgery proposed by the present invention. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The following describes in detail a specific solution of a facial feature recognition and remodeling optimization method for plastic surgery provided by the present invention with reference to the accompanying drawings.

[0020] See also Figure 1 , which shows a facial feature recognition and reshaping optimization method for plastic surgery provided by an embodiment of the present invention, the method comprising: step S1: obtaining expression images of the patient's facial state, the expression images are divided into natural expression images in a natural facial state and special expression images of all other facial states.

[0021] The embodiments of the present invention are primarily used in scenarios involving facial plastic surgery and optimization. In practice, before adjusting a patient's face, relevant personnel must first create a 3D model of the patient's face. This 3D modeling requires a large number of facial expression images of the patient's facial states. Therefore, in the embodiments of the present invention, facial expression images of the patient's facial states are obtained. Because a person's facial structure is most clearly defined when they are in a natural facial expression state, natural expression images of the face in a natural state are obtained for comparison. At the same time, special expression images of various facial forms, such as smiling, frowning, surprised, and sad, are obtained. By comparing the two, different facial muscle movements and changes are highlighted.

[0022] In one embodiment of the present invention, a high-resolution HD camera is used to capture the patient's facial expressions in different states to obtain facial expression images of the patient. It should be noted that the image capture process is a technical means well known to those skilled in the art and will not be limited or elaborated herein.

[0023] Step S2: perform key point detection on all expression images to obtain key points in each expression image; use the key points to divide all expression images into regions to obtain all facial regions in each expression image; select a facial region in any expression image as the target region; obtain the information richness index of the target region based on the key point distribution, edge distribution and distribution characteristics of other adjacent facial regions in the target region; grade the facial regions according to the information richness index to obtain the regional level of each facial region.

[0024] In plastic surgery, different areas of the face have varying importance, but the shape and proportions of each area significantly impact the overall facial beauty and harmony. To better capture the dynamic features of the face and ensure the accuracy and authenticity of plastic surgery, it is necessary to identify and grade key areas in the patient's facial image to obtain more accurate facial features. Therefore, in this embodiment of the present invention, key point detection is performed on all facial images to obtain the key points in each image; and key points are used to divide all facial images into regions to obtain all facial regions in each image.

[0025] Different facial regions contain varying degrees of facial information richness. This allows for ranking different facial regions to determine their importance. This allows for subsequent comparison of the ranking differences between different facial regions in natural expression images and special expression images to identify similar structures within each facial region in the natural expression image and in the other special expression images. Therefore, in this embodiment of the present invention, an information richness index for the target region is derived based on the distribution of key points, edge distribution, and other adjacent facial regions within the target region, allowing for the classification of the facial regions.

[0026] Preferably, in one embodiment of the present invention, the method for obtaining the information richness index includes: obtaining the information richness index according to an information richness index calculation formula, the information richness index calculation formula is as follows: Where, Information richness indicator representing the target area; Indicates the number of key points in the target area; Indicates the area of the target region; Indicates the number of edge pixels in the target area; Indicates the The gradient value of edge pixels; Indicates the number of facial regions adjacent to the target region.

[0027] In the information richness index calculation formula, the density of the number of key points in the target area The larger the value, the higher the information richness in the target area, that is, the larger the information richness index of the target area; the average gradient of the edge pixels of the target area The larger the value is, the more significant the boundary of the target area is, and the higher the information richness in the target area is, that is, the larger the information richness index of the target area is; the more facial areas adjacent to the target area are, the more obvious the gradient characteristics of the target area are, and the higher the information richness in the target area is, that is, the larger the information richness index of the target area is.

[0028] Preferably, in one embodiment of the present invention, the method for obtaining the region level of each facial region includes: setting a preset first number of preset threshold ranges; sorting all the preset threshold ranges from small to large, and marking a serial number for each preset threshold range as the region level to which each preset threshold range belongs; in one embodiment of the present invention, the preset threshold ranges are respectively set to , marked as level 1; , marked as level 2; , marked as level 3; , marked as level 4; , marked as level 5; , marked as level 6; , marked as level 7; , marked as level 8; , marked as level 9; , marked as level 10. It should be noted that the number of preset threshold ranges and regional levels can be set arbitrarily and are not limited here.

[0029] Each facial region in the expression image is divided into its own preset threshold range according to the information richness index to obtain the regional level of each facial region.

[0030] Step S3: randomly select a key point in the natural expression image as a reference point; obtain the corresponding point of the reference point in each special expression image based on the similarity features between the special expression image and the natural expression image; obtain the basic stability factor of the reference point in each special expression image based on the distance features between the reference point and the corresponding point and the facial area differences; obtain the structural similarity of the facial area where the reference point is located in each special expression image based on the distance features, area level differences and grayscale feature differences within the area between the facial area where the reference point is located and the facial area of each special expression image; obtain the final stability factor of the reference point based on the basic stability factor and the structural similarity.

[0031] Facial muscle movements under different expressions vary from person to person, and each person's facial expressions and dynamic changes are subtly different. To make the constructed facial model more accurate and personalized, it is necessary to extract variation characteristics by analyzing the variation patterns of each facial region under different expressions. Knowing that the positions and numbers of key points detected under different facial expressions can vary significantly, it is difficult to match all key points one by one. Therefore, we analyze the facial regions where the key points are located and calculate the corresponding stability factors, which serve as the parameter basis for constructing the facial model.

[0032] Preferably, in one embodiment of the present invention, the method for obtaining the corresponding point of the reference point in each special expression image includes: selecting any special expression image as a comparison image; marking the corresponding position of the reference point in the comparison image, and taking the key point in the comparison image closest to the corresponding position of the reference point as the corresponding point of the reference point.

[0033] Preferably, in one embodiment of the present invention, the method for obtaining the basic stability factor includes: obtaining the basic stability factor according to a basic stability factor calculation formula, and the basic stability factor calculation formula is as follows: Where, Indicates the basic stability factor of the reference point in each special expression image; Indicates the distance between the corresponding position of the reference point and the position of the comparison point in each special expression image; Represents the area of overlap between the facial region to which the reference point belongs and the facial region to which the corresponding point belongs; The area of the largest facial region between the reference point and the corresponding point; Represents an exponential function with a natural constant as its base.

[0034] In the basic stability factor calculation formula, the smaller the distance between the corresponding position of the reference point in each special expression image and the position of the comparison point, the more stable the position of the reference point is under the change of expression. The larger the value, the greater the basic stability factor; the ratio of the overlapping area to the larger facial area The larger it is, the greater the area similarity and overlap between the facial area to which the reference point belongs and the facial area to which the corresponding point belongs, which means that the facial area where the reference point is located is more stable under changes in expression, and the greater the basic stability factor of the reference point in each special expression image.

[0035] Because the magnitude of facial variation for each unnatural expression is different from that for natural expressions, and the magnitude of their impact on key points is also different, the final stability factor cannot simply be calculated by combining all basic stability factors. The magnitude of variation for each expression must also be considered to prevent certain unique expressions from affecting the accuracy of the stability factor. Therefore, in this embodiment of the present invention, the degree of structural similarity of the facial region where the reference point is located in each special expression image is determined based on the distance characteristics, regional level differences, and grayscale feature differences between the facial region where the reference point is located and the facial region in each special expression image.

[0036] Preferably, in one embodiment of the present invention, the method for obtaining the degree of structural similarity includes: obtaining the corresponding area of the area to which the reference point belongs in the comparison image; selecting a preset second number of facial areas closest to the area to which the reference point belongs as the neighboring areas of the facial area where the reference point is located; selecting a preset second number of facial areas closest to the area to which the corresponding point belongs as the neighboring areas of the facial area where the corresponding point is located; in one embodiment of the present invention, the preset second number is set to 5. It should be noted that, in other embodiments of the present invention, the preset second number can be set arbitrarily and is not limited here.

[0037] The structural similarity is obtained according to the structural similarity calculation formula. The structural similarity calculation formula is as follows: Where, Indicates the structural similarity of the facial region where the reference point is located in each special expression image; Indicates the number of neighboring regions of the facial region where the reference point is located; Indicates the facial region where the reference point is located. The center pixel of the neighboring area and the corresponding point in the facial area The distance between the center pixels of adjacent areas; Indicates the facial region where the reference point is located. The neighboring area and the facial area where the corresponding point is located Differences in regional levels between adjacent regions; Indicates the grayscale mean difference between the facial area where the reference point is located and the facial area where the corresponding point is located; represents the absolute value function; Represents an exponential function with a natural constant as its base.

[0038] In the structural similarity calculation formula, the first The center pixel of the neighboring area and the corresponding point in the facial area The smaller the distance between the center pixels of the adjacent regions, the higher the position similarity between the two facial regions. The smaller the regional level difference between the two facial regions, the higher the information similarity between the two facial regions. Each adjacent region is analyzed, that is, The smaller the value, the smaller the structural difference between the reference point and the corresponding point in the surrounding area. At this time, the structural similarity of the reference point is greater. The absolute value of the average grayscale difference between the reference point and the corresponding point is The smaller it is, the smaller the light intensity at the corresponding position is, that is, the facial structure changes in the same facial area after the expression changes are smaller, and the structural similarity of the facial area where the reference point is located is greater.

[0039] Preferably, in one embodiment of the present invention, the method for obtaining the final stability factor includes: obtaining the final stability factor according to a final stability factor calculation formula, and the final stability factor calculation formula is as follows: Where, represents the final stability factor of the reference point; Indicates the number of special expression images; Indicates the reference point Basic stability factor of special expression images; Indicates the facial area where the reference point is located in The degree of structural similarity in the special expression images; Represents the normalization function.

[0040] In the final stability factor calculation formula, the basic stability factor is weighted using the degree of structural similarity. When the basic stability factors of the reference points are the same, the greater the local change amplitude, the more it can reflect its stability. Therefore, the lower the degree of structural similarity, the higher the weight of the basic stability factor.

[0041] Step S4: Obtaining an optimized facial model of the patient according to the final stabilization factor; and performing plastic surgery optimization on the patient's face according to the optimized facial model.

[0042] In the prior art, in the process of constructing a facial model based on a natural expression image, it is necessary to select multiple 3D facial models from a template database based on key points, and perform shape interpolation through multiple 3D models to generate a final three-dimensional facial model. In the process of selecting a template model, in order to screen out the most suitable model, a common method is to minimize the reprojection error, that is, after projecting the 3D template onto the expression image, it is compared with the key points to minimize the error. In actual scenarios, key points at different locations on the face have different reference values. Therefore, the embodiment of the present invention combines the changing characteristics of the key points, uses the final stability factor obtained in the previous step, and recalculates the reprojection error, so that the screened 3D template model is more consistent with the dynamic characteristics of the patient's face, and enhances the structural integrity of the final model.

[0043] In the embodiment of the present invention, the calculation formula of the total reprojection error is as follows: Where, represents the total reprojection error between the natural expression image and each 3D model; Indicates the number of key points in natural expression images; Indicates the first The reprojection error of key points can be directly obtained by existing technology; Indicates the first The final stabilization factor of a key point.

[0044] In the calculation formula of the total reprojection error, The larger the final stability factor of a key point is, the greater the error weight of the key point is. Each key point in the natural expression image is analyzed to obtain the total reprojection error between the expression image and each 3D model.

[0045] Filter out the 3 with the smallest total reprojection error 50 template models, the specific number is determined by the required accuracy. Finally, shape interpolation is completed based on the selected models to generate a 3D facial model.

[0046] After completing the reshaping of the 3D facial model, relevant personnel can conduct professional evaluation and feedback based on the optimized facial model to develop a personalized plastic surgery plan while maximizing the naturalness and harmony of the surgical effect.

[0047] In summary, the facial expression images of the patient's facial state are obtained, and the facial expression images are divided into natural expression images in natural facial states and special expression images of all other facial states; key point detection is performed on all expression images to obtain key points in each expression image; all expression images are divided into regions using key points to obtain all facial regions in each expression image; a facial region in any expression image is selected as the target region; the information richness index of the target region is obtained based on the key point distribution, edge distribution and distribution characteristics of other adjacent facial regions in the target region; the facial regions are graded based on the information richness index to obtain the regional level of each facial region; a key point in the natural expression image is selected as a reference point; according to the similarity features between the special expression images and the natural expression images, the corresponding point of the reference point in each special expression image is obtained; according to the distance features between the reference point and the corresponding point and the difference in facial regions, the basic stability factor of the reference point in each special expression image is obtained; according to the distance features, regional level differences and grayscale feature differences between the facial region where the reference point is located and the facial region of each special expression image, the structural similarity of the facial region where the reference point is located in each special expression image is obtained; according to the basic stability factor and the structural similarity, the final stability factor of the reference point is obtained; according to the final stability factor, the optimized facial modeling of the patient is obtained; and according to the optimized facial modeling, the patient's face is optimized.

[0048] The second purpose of an embodiment of the present invention is to provide a facial feature recognition and reshaping optimization system for plastic surgery, the system comprising a memory, a processor, and a computer program, wherein the memory is used to store the corresponding computer program, the processor is used to run the corresponding computer program, and when the computer program runs in the processor, it can implement the method described in steps S1-S4.

[0049] The third object of an embodiment of the present invention is to provide a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in steps S1-S4 when executing the computer program.

[0050] A fourth object of an embodiment of the present invention is to provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in steps S1-S4 is implemented.

[0051] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0052] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A facial feature recognition and reshaping optimization method for plastic surgery, characterized in that: The method comprises the following steps: obtaining facial expression images of a patient's facial state, wherein the facial expression images are divided into natural facial expression images in a natural facial state and special facial expression images in all other facial states; performing key point detection on all facial expression images to obtain key points in each facial expression image; performing regional division on all facial expression images using the key points to obtain all facial regions in each facial expression image; selecting a facial region in an expression image as a target region; obtaining an information richness index of the target region based on key point distribution, edge distribution and distribution characteristics of other adjacent facial regions in the target region; grading the facial regions based on the information richness index to obtain a region grade of each facial region; selecting a key point in a natural facial image as a reference region; Test points: According to the similarity features between special expression images and natural expression images, obtain the corresponding point of the reference point in each special expression image; according to the distance features between the reference point and the corresponding point and the difference in facial areas, obtain the basic stability factor of the reference point in each special expression image; according to the distance features, area level differences and grayscale feature differences between the facial area where the reference point is located and the facial area of each special expression image, obtain the structural similarity of the facial area where the reference point is located in each special expression image; according to the basic stability factor and the structural similarity, obtain the final stability factor of the reference point; according to the final stability factor, obtain the optimized facial modeling of the patient; according to the optimized facial modeling, perform plastic surgery on the patient's face.

2. A facial feature recognition and reshaping optimization method for plastic surgery according to claim 1, characterized in that: The method for obtaining the information richness index includes: obtaining the information richness index according to an information richness index calculation formula, and the information richness index calculation formula is as follows: Where, Information richness indicator representing the target area; Indicates the number of key points in the target area; Indicates the area of the target region; Indicates the number of edge pixels in the target area; Indicates the The gradient value of edge pixels; Indicates the number of facial regions adjacent to the target region.

3. The facial feature recognition and reshaping optimization method for plastic surgery according to claim 1, characterized in that: The method for obtaining the regional level of each facial area includes: setting a preset first number of preset threshold ranges; sorting all the preset threshold ranges from small to large, and marking each preset threshold range with a serial number as the regional level to which each preset threshold range belongs; and dividing each facial area in the expression image into the preset threshold range to which it belongs according to the information richness index, thereby obtaining the regional level of each facial area.

4. The facial feature recognition and reshaping optimization method for plastic surgery according to claim 1, characterized in that: The method for obtaining the corresponding point of the reference point in each special expression image includes: selecting any special expression image as a comparison image; marking the corresponding position of the reference point in the comparison image, and taking the key point closest to the corresponding position of the reference point in the comparison image as the corresponding point of the reference point.

5. The facial feature recognition and reshaping optimization method for plastic surgery according to claim 1, characterized in that: The method for obtaining the basic stability factor includes: obtaining the basic stability factor according to a basic stability factor calculation formula, and the basic stability factor calculation formula is as follows: Where, Indicates the basic stability factor of the reference point in each special expression image; Indicates the distance between the corresponding position of the reference point and the position of the comparison point in each special expression image; Represents the area of overlap between the facial region to which the reference point belongs and the facial region to which the corresponding point belongs; The area of the largest facial region between the reference point and the corresponding point; Represents an exponential function with a natural constant as its base.

6. The facial feature recognition and reshaping optimization method for plastic surgery according to claim 1, characterized in that: The method for obtaining the degree of structural similarity includes: obtaining a corresponding area of the area to which the reference point belongs in the comparison image; selecting a preset second number of facial areas closest to the area to which the reference point belongs as neighboring areas of the facial area where the reference point is located; selecting a preset second number of facial areas closest to the area to which the corresponding point belongs as neighboring areas of the facial area where the corresponding point is located; and obtaining the degree of structural similarity according to a structural similarity calculation formula, wherein the structural similarity calculation formula is as follows: Where, Indicates the structural similarity of the facial region where the reference point is located in each special expression image; Indicates the number of neighboring regions of the facial region where the reference point is located; Indicates the facial region where the reference point is located. The center pixel of the neighboring area and the corresponding point in the facial area The distance between the center pixels of adjacent areas; Indicates the facial region where the reference point is located. The neighboring area and the facial area where the corresponding point is located Differences in regional levels between adjacent regions; Indicates the grayscale mean difference between the facial area where the reference point is located and the facial area where the corresponding point is located; represents the absolute value function; Represents an exponential function with a natural constant as its base.

7. The facial feature recognition and reshaping optimization method for plastic surgery according to claim 1, characterized in that: The method for obtaining the final stability factor includes: obtaining the final stability factor according to a final stability factor calculation formula, and the final stability factor calculation formula is as follows: Where, represents the final stability factor of the reference point; Indicates the number of special expression images; Indicates the reference point Basic stability factor of special expression images; Indicates the facial area where the reference point is located in The degree of structural similarity in the special expression images; Represents the normalization function.

8. A facial feature recognition and reshaping optimization system for plastic surgery, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the facial feature recognition and reshaping optimization method for plastic surgery as described in any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the facial feature recognition and reshaping optimization method for plastic surgery as claimed in any one of claims 1 to 7 are implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the facial feature recognition and reshaping optimization method for plastic surgery as claimed in any one of claims 1 to 7 are implemented.

Images