Face morphing recommendation method and device based on image recognition, equipment and medium

Abstract

The application relates to the technical field of artificial intelligence, and provides a face deformation recommendation method and device based on image recognition, equipment and a medium. The method comprises the following steps: displaying a plurality of selectable face styles; in response to a style selection instruction of at least one face style in the plurality of face styles, comparing face features corresponding to a face image to be deformed provided by a user with face common features corresponding to the selected face style, obtaining face deformation recommendation information corresponding to a face feature comparison result, and displaying the face deformation recommendation information and a face deformation effect obtained by applying the face deformation recommendation information to the face image to be deformed. Therefore, the user can be provided with rich selectable face styles, the face features corresponding to the face image to be deformed provided by the user can be compared with the face common features corresponding to the selected style according to the selected style, and the user can be objectively and accurately provided with face deformation recommendation information and face deformation effect preview corresponding to the comparison result.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a facial deformation recommendation method, apparatus, computer equipment, and storage medium based on image recognition. Background Art

[0002] With the development of artificial intelligence, facial image recognition technology has emerged. This technology can identify facial features contained in images and is widely used in face-related scenarios. In particular, applying facial image recognition technology to face morphing recommendations facilitates providing users with relevant face morphing suggestions.

[0003] However, the facial deformation recommendation method provided by current technology needs to rely on the facial deformation threshold set by a specific user group based on their subjective standards to recommend facial deformation schemes to users. Moreover, the standard facial deformation threshold set by this method is often too simple and cannot objectively and accurately recommend facial deformation schemes to users. Summary of the Invention

[0004] Based on this, it is necessary to provide a facial deformation recommendation method, device, computer equipment and storage medium based on image recognition to address the above technical problems.

[0005] A facial deformation recommendation method based on image recognition, the method comprising:

[0006] Showcase a variety of facial styles;

[0007] In response to a style selection instruction for at least one of the plurality of facial styles, obtaining facial deformation recommendation information corresponding to a facial feature comparison result obtained by comparing facial features corresponding to a facial image to be deformed provided by a user with common facial features corresponding to the selected facial style;

[0008] The facial deformation recommendation information is displayed, and a facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed is displayed.

[0009] A facial deformation recommendation device based on image recognition, comprising:

[0010] Style display module, used to display multiple facial styles;

[0011] an information acquisition module configured to, in response to a style selection instruction for at least one of the plurality of facial styles, acquire facial deformation recommendation information corresponding to a facial feature comparison result obtained by comparing facial features corresponding to a facial image to be deformed provided by a user with common facial features corresponding to the selected facial style;

[0012] The deformation display module is used to display the facial deformation recommendation information and the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed.

[0013] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0014] Display multiple facial styles; in response to a style selection instruction for at least one of the multiple facial styles, obtain facial deformation recommendation information corresponding to the facial feature comparison result; the facial feature comparison result is obtained by comparing the facial features corresponding to the facial image to be deformed provided by the user with the facial common features corresponding to the selected facial style; display the facial deformation recommendation information, and display the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed.

[0015] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0016] Display multiple facial styles; in response to a style selection instruction for at least one of the multiple facial styles, obtain facial deformation recommendation information corresponding to the facial feature comparison result; the facial feature comparison result is obtained by comparing the facial features corresponding to the facial image to be deformed provided by the user with the facial common features corresponding to the selected facial style; display the facial deformation recommendation information, and display the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed.

[0017] A computer program product or computer program includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps of the above method.

[0018] The aforementioned image recognition-based facial deformation recommendation method, apparatus, computer device, and storage medium display a variety of optional facial styles. In response to a user's selection of at least one of the multiple facial styles, the method compares the facial features corresponding to the user-provided facial image to be deformed with the common facial features corresponding to the selected facial style, obtains facial deformation recommendation information corresponding to the facial feature comparison result, and then displays the facial deformation recommendation information and the facial deformation effect obtained by applying the facial deformation recommendation information to the aforementioned facial image to be deformed. This solution can provide users with a rich selection of facial styles and, based on the user's selected style, compares the facial features corresponding to the user's provided facial image to be deformed with the common facial features corresponding to the selected style, thereby objectively and accurately providing users with facial deformation recommendation information corresponding to the comparison result and a preview of the facial deformation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 FIG. 1 is a diagram illustrating an application environment of a facial deformation recommendation method based on image recognition in one embodiment;

[0020] Figure 2 1 is a flow chart of a facial deformation recommendation method based on image recognition in one embodiment;

[0021] Figure 3 This is a schematic diagram of an interface showing facial styles in one embodiment;

[0022] Figure 4 A flowchart illustrating the steps of obtaining facial deformation recommendation information corresponding to facial feature comparison results in one embodiment;

[0023] Figure 5 A schematic diagram of an interface for displaying a facial template image represented by common facial features in one embodiment;

[0024] Figure 6 A schematic diagram of an interface for importing a facial image to be deformed in one embodiment;

[0025] Figure 7 A schematic diagram of an interface displaying facial deformation recommendation information in one embodiment;

[0026] Figure 8 A schematic diagram of an interface for previewing facial deformation effects in one embodiment;

[0027] Figure 9 A flowchart illustrating the steps of obtaining common facial features corresponding to various facial styles in one embodiment;

[0028] Figure 10 A flowchart illustrating the steps of determining common facial features of multiple facial image samples in one embodiment;

[0029] Figure 11 This is a flowchart of a method for intelligently recommending cosmetic surgery plans and previewing results based on facial recognition algorithms and deep learning in an application example.

[0030] Figure 12 This is a schematic diagram of the structure of a cosmetic surgery plan intelligent recommendation and effect preview system based on facial recognition algorithm and deep learning in an application example;

[0031] Figure 13 Generate a data processing flow chart for the module of the system solution in an application example;

[0032] Figure 14 is a structural block diagram of a facial deformation recommendation device based on image recognition in one embodiment;

[0033] Figure 15 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0035] The face deformation recommendation method based on image recognition provided by this application can be applied to Figure 1 In the application environment shown, the terminal 110 can communicate with the server 120 via a network. The terminal 110 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices, and the server 120 can be implemented as an independent server or a server cluster consisting of multiple servers.

[0036] Specifically, the facial deformation recommendation method based on image recognition provided in this application can be executed by terminal 110 alone, or by terminal 110 in conjunction with server 120. In the case of execution by terminal 110 alone, terminal 110 displays multiple facial styles to the user, and the user can select at least one of the facial styles on terminal 110. The selected at least one facial style is referred to as the selected facial style. Then, in response to the user's style selection instruction for at least one of the displayed multiple facial styles, terminal 110 obtains facial deformation recommendation information corresponding to a facial feature comparison result. The facial feature comparison result can be obtained by terminal 110 based on a comparison of facial features corresponding to the facial image to be deformed, provided by the user, with common facial features corresponding to the selected facial style. After obtaining the facial feature comparison result, terminal 110 obtains facial deformation recommendation information corresponding to the facial feature comparison result, displays the facial deformation recommendation information to the user, and displays the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed.

[0037] In addition, in a scenario where the terminal 110 and the server 120 cooperate to perform the function, the terminal 110 may primarily perform the displayed steps, and the server 120 may provide the terminal 110 with the facial feature comparison process and facial style information. Specifically, the server 120 may, using artificial intelligence technology, analyze and learn facial images of multiple facial styles to determine the common features shared by each facial image under each facial style, referred to as common facial features. These common facial features may be pre-stored and associated with the corresponding facial style. In other words, the server 120 may be able to find the corresponding common facial features based on the facial style.

[0038] Therefore, the server 120 can provide the terminal 110 with a variety of facial styles for the user to choose from based on the facial image learning results based on artificial intelligence technology. When the user needs facial deformation recommendations, the terminal 110 displays the multiple facial styles. The number of the multiple facial styles can be continuously updated according to the server 120's learning of facial images, so that the facial styles available for users to choose can be continuously enriched. After terminal 110 displays multiple facial styles, the user can select at least one of the facial styles on terminal 110. Terminal 110 then requests server 120 to obtain facial deformation recommendation information corresponding to the facial feature comparison results in response to the user's style selection instruction for at least one of the multiple facial styles. That is, in a scenario where terminal 110 and server 120 cooperate with each other, the facial feature comparison process and the process of obtaining facial deformation recommendation information can be handed over to server 120 to reduce the data processing pressure of terminal 110. Similarly, server 120 can compare the facial features corresponding to the facial image to be deformed provided by the user with the common facial features corresponding to the selected facial style for terminal 110, thereby obtaining facial deformation recommendation information corresponding to the facial feature comparison results, as well as the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed. At this point, server 120 can feed back the facial deformation recommendation information and the corresponding facial deformation effect to terminal 110, and terminal 110 displays the received facial deformation recommendation information and facial deformation effect to the user.

[0039] It can be understood that the process of deriving common facial features based on learning and analysis of facial images of various facial styles with the help of artificial intelligence technology in the scenario executed by the terminal 110 and the server 120 as above, and the process of performing feature comparison and providing facial deformation recommendation information and facial deformation effects after the user selects a style, can also be fully or partially executed by the terminal 110 in the scenario where the terminal 110 alone executes the method provided in this application.

[0040] The above scenario can reflect the facial deformation recommendation method based on image recognition provided by this application, which can be realized with the help of artificial intelligence technology, so that this application can provide users with objective and accurate facial deformation recommendation information and corresponding facial deformation effect preview on the basis of intelligence.

[0041] Artificial Intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also involves studying the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0042] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0043] Computer vision (CV) is the science of making machines "see." Specifically, it refers to machine vision, where cameras and computers replace the human eye in identifying and measuring objects, performing further image processing to create images more suitable for human observation or transmission to instruments. As a scientific discipline, computer vision studies related theories and technologies, attempting to build artificial intelligence systems capable of extracting information from images or multidimensional data. Computer vision technologies typically include image processing, image recognition, image semantic understanding, image retrieval, optical character recognition (OCR), video processing, video semantic understanding, video content / behavior recognition, three-dimensional object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping, and common biometric recognition technologies such as facial recognition and fingerprint recognition.

[0044] Machine learning (ML) is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. It specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of AI. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.

[0045] Based on this, the facial deformation recommendation method based on image recognition provided by this application can use computer vision technology and machine learning in artificial intelligence technology to identify and learn the facial feature information of a large number of facial images, and analyze the common facial features corresponding to different facial styles, so that after the user provides the facial image to be deformed and confirms the selected facial style, the facial feature differences with the selected facial style are analyzed and corresponding facial deformation recommendation information and facial deformation effect preview are provided. This can avoid the problem that the facial deformation recommendation effect is single, subjective and not accurate due to the traditional technology relying on the facial deformation threshold set by a specific user group based on its subjective standards, and achieve the effect of objectively and accurately providing users with facial deformation recommendation information and facial deformation effect preview corresponding to the comparison results.

[0046] The following is a further description of the facial deformation recommendation method based on image recognition provided by the present application in conjunction with the embodiments and corresponding drawings.

[0047] In one embodiment, Figure 2 As shown in the figure, a facial deformation recommendation method based on image recognition is provided. Figure 1 The terminal 110 in FIG. 1 is taken as an example to illustrate the process, including the following steps:

[0048] Step S201, displaying multiple facial styles;

[0049] In this step, if Figure 3 As shown, after terminal 110 enters the facial morphing recommendation system, a plurality of facial styles may be displayed on first interface 310 for the user to select. The user may select one or more of these styles. To facilitate the user's selection of a facial style, terminal 110 may also display a facial template image corresponding to the facial style near the corresponding facial style position on first interface 310. For example, the user may select at least one of facial styles 1 to 5 on first interface 310. The user may also select a custom facial style on first interface 310 to meet the user's personalized needs for facial morphing.

[0050] Among them, facial style refers to the style of the face of an object mainly represented by a person with attributes such as different regions, identities, ages and / or genders. That is, the faces of people with different regions, identities, ages and / or genders usually have different styles, while the faces of people with the same regions, identities, ages and / or genders usually have certain commonalities.

[0051] Thus, terminal 110 can provide multiple facial styles and display facial template images corresponding to each facial style for user reference and selection. For example, the facial styles provided by terminal 110 can be, for example, regional style A, regional style B, and a mixed style of regional styles AB, etc. The facial styles provided by terminal 110 can also include custom styles provided by the user. The custom style is a facial style obtained by terminal 110 or server 120 using artificial intelligence technology to analyze multiple facial images provided by the user. The user can refer to the corresponding facial template images and select one or more facial styles from them, and then submit them to terminal 110 for subsequent recommendation. In the subsequent process description, the process of facial deformation recommendation is mainly explained by selecting a facial style as an example.

[0052] Step S202 : In response to a style selection instruction for at least one facial style from among a plurality of facial styles, facial deformation recommendation information corresponding to a facial feature comparison result is obtained.

[0053] Specifically, the user can trigger a style selection instruction for the facial style by clicking on the facial template image corresponding to the desired facial style (referred to as the selected facial style) on the first interface 310 provided by the terminal 110. In response to the style selection instruction, the terminal 110 can compare the facial features corresponding to the facial image to be deformed provided by the user with the common facial features corresponding to the selected facial style. The result of the feature comparison is called the facial feature comparison result, and facial deformation recommendation information corresponding to the facial feature comparison result is obtained.

[0054] Before the terminal 110 performs the facial feature comparison process, the user is required to provide a facial image to be deformed. The facial image to be deformed refers to an image containing the face to be deformed, which can be the user's own facial image or the facial image of another user. After obtaining the facial image to be deformed provided by the user and determining the facial style selected by the user, the terminal 110 can analyze the facial features of the facial image to be deformed based on the facial recognition algorithm and compare it with the common facial features corresponding to the facial style selected by the user. The facial feature comparison result obtained by the terminal 110 can be used to characterize the main difference between the facial features of the facial image to be deformed and the common facial features corresponding to the selected facial style. The terminal 110 can give facial deformation recommendation information based on the main difference, such as raising the nose bridge by 1mm, etc., that is, facial deformation recommendation information, which refers to information recommending what kind of deformation of the face presented in the facial image to be deformed in order to achieve the selected facial style.

[0055] In addition to obtaining facial deformation recommendation information, terminal 110 can further obtain the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed. The facial deformation effect can be presented in a static or dynamic form. The static form can be an image obtained by applying the facial deformation recommendation information to the facial image to be deformed, while the dynamic form can be a process from the image to be deformed to the image obtained after applying the facial deformation recommendation information to the facial image to be deformed in the form of a video or animated image.

[0056] Step S203 : displaying the facial deformation recommendation information and the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed.

[0057] This step mainly involves the terminal 110 obtaining the facial deformation recommendation information and applying the facial deformation recommendation information to the facial image to be deformed to obtain the facial deformation effect, and then displaying the facial deformation recommendation information and the facial deformation effect to the user.

[0058] In the above-mentioned facial deformation recommendation method based on image recognition, terminal 110 displays a variety of optional facial styles. In response to the user's style selection instruction for at least one of the multiple facial styles, terminal 110 can compare the facial features corresponding to the facial image to be deformed provided by the user with the facial common features corresponding to the selected facial style, obtain facial deformation recommendation information corresponding to the facial feature comparison result, and then terminal 110 displays the facial deformation recommendation information and the facial deformation effect obtained by applying the facial deformation recommendation information to the aforementioned facial image to be deformed. This solution can provide users with a rich variety of optional facial styles and compare the facial features corresponding to the facial image to be deformed provided by the user with the facial common features corresponding to the selected style based on the style selected by the user, thereby objectively and accurately providing users with facial deformation recommendation information corresponding to the comparison result and a preview of the facial deformation effect.

[0059] In some embodiments, before performing facial feature comparison for the user, the terminal 110 allows the user to adjust the common facial features corresponding to the selected facial style, so that the facial deformation recommendation is more personalized. Figure 4 As shown, in step S202, in response to a style selection instruction for at least one facial style from a plurality of facial styles, obtaining facial deformation recommendation information corresponding to the facial feature comparison result may include the following steps:

[0060] Step S401, in response to a style selection instruction, displaying a facial template image represented by common facial features, and adaptively changing the display of the facial template image when the common facial features are adjusted by the user;

[0061] refer to Figure 5After the user selects one of the facial styles from a variety of facial styles, for example, the terminal 110 responds to the style selection instruction and displays a facial template image 511 represented by the common facial features K1 corresponding to the selected facial style on the second interface 510. It should be noted that the common facial features K1 displayed on the second interface 510 mainly illustrate a set of partial feature points located on the facial contour, and the common facial features K1 corresponding to the facial style are not limited to a set of feature points located on the facial contour, but also include a set of feature points composed of feature points located in other parts such as the eyes, mouth, eyebrows and nose. In order to facilitate the description of the main content of the solution, only the set of partial feature points located on the facial contour is illustrated in the figure. In actual application, if the user is satisfied with the selected facial style, the facial style can be marked in accordance with the satisfaction level in the satisfaction selection area 512.

[0062] Furthermore, the user can adjust the common facial features, and the terminal 110 adaptively changes the display of the facial template image when the common facial features are adjusted by the user. Specifically, a facial template image 511 represented by the common facial feature K1 is displayed on the second interface 510. The user can adjust the position of one or more feature points in the common facial feature K1. When the position of one or more feature points is adjusted, the terminal 110 can change the display of the facial template image for the user in real time based on the adjusted position of the corresponding feature point. That is, during the display of the facial template image, the user can adjust and modify the selected facial style and preview the modification effect in real time.

[0063] Specifically, the user's self-adjustment of the selected facial style may be achieved through the following steps. In one embodiment, the adaptive change of the display of the facial template image when the common facial features are adjusted by the user in step S401 may include:

[0064] Provide common facial features corresponding to the common facial feature point position parameters on the facial template image, and adaptively change the display of the facial template image according to the user's adjustment of the common facial feature point position parameters.

[0065] Still refer to Figure 5, the terminal 110 may display a parameter adjustment area 513 on the second interface 510, in which the position parameters (or position coordinates) of the facial common feature points corresponding to the facial common feature K1 of the selected facial style on the facial template image 511 are provided. Thus, the user can change the position coordinates (such as x1, y1; x2, ...) of one or more facial common feature points in the parameter adjustment area 513, so that the terminal 110 adaptively changes the display of the facial template image 511 according to the user's adjustment of the facial common feature point position parameters in the parameter adjustment area 513, thereby assisting the user to conveniently perform personalized adjustments and modifications to the facial common features after selecting the corresponding facial style and providing a real-time preview of the modification effect.

[0066] Step S402 : In response to a facial feature comparison instruction from a user, facial deformation recommendation information corresponding to the facial feature comparison result is obtained.

[0067] like Figure 5 As shown, after browsing the facial template image corresponding to the selected facial style, the user can trigger a facial feature comparison instruction in the "Compare with Me" option on the second interface 510. In response to the facial feature comparison instruction, after the user provides the facial image to be deformed, the terminal 110 can compare the facial features of the facial image to be deformed with the common facial features to obtain corresponding facial deformation recommendation information. Since the user can choose to adjust or not modify the common facial features corresponding to the selected facial style on the second interface 510, if the user does not adjust the common facial features corresponding to the selected facial style, the facial feature comparison result obtained by the terminal 110 is the comparison result of the common facial features not adjusted by the user and the facial features corresponding to the facial image to be deformed. If the user adjusts the common facial features, the facial feature comparison result obtained by the terminal 110 is the comparison result of the common facial features adjusted by the user and the facial features corresponding to the facial image to be deformed.

[0068] Among them, the terminal 110 can support multiple ways for the user to provide the facial image to be deformed, for example, it can use the terminal 110 camera to collect the user's multi-angle facial information in real time or import images from the terminal 110's image library, etc., to meet the user's facial information collection needs when it is inconvenient to turn on the camera. Specifically, in some embodiments, the user can Figure 5 After clicking "Compare with Me" on the second interface 510 shown in FIG. 1 , the terminal 110 may display the following Figure 6The third interface 610 shown in the figure allows the user to use a camera or import a facial image to be deformed from the image library of the terminal 110. After the user imports the facial image to be deformed, he can click "Start Comparison" to instruct the terminal 110 to perform a comparison process between the facial features corresponding to the facial image to be deformed and the facial common features corresponding to the selected facial style, so that the terminal 110 obtains the corresponding facial deformation recommendation information based on the facial feature comparison results.

[0069] like Figure 7 As shown, after obtaining the facial deformation recommendation information, the terminal 110 can display it on the fourth interface 710. In some embodiments, when the terminal 110 displays the facial deformation recommendation information 714 on the fourth interface 710, the user can modify and adjust the facial deformation recommendation information 714 and click "Effect Preview" to view the preview effect (that is, displaying the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed).

[0070] Specifically, the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed in step S203 may include:

[0071] According to the user's adjustment of the facial deformation recommendation information, the adjusted facial deformation recommendation information is obtained; in response to the effect preview instruction from the user, the facial deformation effect obtained by applying the adjusted facial deformation recommendation information to the facial image to be deformed is displayed.

[0072] In this embodiment, the user Figure 6 After clicking "Start Comparison" on the third interface 610 to trigger the terminal 110 to perform the facial feature comparison process, the terminal 110 obtains the facial deformation recommendation information corresponding to the facial feature comparison result and provides the following information: Figure 7 In the fourth interface 710 shown, the terminal 110 displays a facial template image 711 represented by a common facial feature K1 on the fourth interface 710, and displays a facial image 712 to be deformed and its corresponding facial feature K2 provided by the user, and displays facial deformation recommendation information corresponding to the facial feature comparison result in the recommendation information display area 714. If the user is satisfied with the comparison process and result of the facial deformation recommendation information, the user can mark / collect the comparison process and result of the facial deformation recommendation information in the comparison result marking / collection area 713.

[0073] Furthermore, the terminal 110 provides the recommended facial deformation information in the recommended information display area 714 of the fourth interface 710, which can be modified and adjusted by the user. Specifically, the user can modify the parameters of each item in the recommended facial deformation information, and even delete or add some adjustment items. For example, the user can modify the "2mm" in "Nose bridge raised by 2mm", or even delete "Nose bridge raised by 2mm", or add other adjustment items. This allows the terminal 110 to provide users with greater autonomy and customization capabilities in the fourth interface 710, avoiding the defect of facial deformation effects that are monotonous and lack personalization.

[0074] After the user adjusts the facial deformation recommendation information, the user can click "Effect Preview" to trigger the terminal 110 to apply the adjusted facial deformation recommendation information to the face image to be deformed 712 to obtain the facial deformation effect, and then Figure 8 The fifth interface 810 shown displays the facial deformation effect, which can be a picture. The user can save the preview picture. That is, after the user adjusts the facial deformation recommendation plan by himself, he can view the expected facial deformation effect of the recommendation plan on the facial image to be deformed 712, thereby supporting the user to adjust the facial deformation recommendation plan more accurately to achieve the expected facial shape effect.

[0075] In one embodiment, Figure 9 As shown, before displaying multiple facial styles in step S201, the above method can also use the following steps to obtain common facial features corresponding to each facial style, specifically including:

[0076] Step S901: Obtain a facial image sample set corresponding to each facial style.

[0077] In this step, the terminal 110 can obtain facial image sample sets corresponding to various facial styles. The facial image sample sets corresponding to each facial style can include multiple facial image samples that meet preset recommendation conditions, that is, the multiple facial image samples contained in each facial image sample set all meet the preset recommendation conditions. The preset recommendation conditions may include, for example, conditions having preset identity attributes and preset age attributes, and may also include conditions imported by the user, that is, images of the faces of objects that meet the preset identity attributes and preset age attributes can be included in the facial image sample set, and images imported by the user can also be included in the facial image sample set.

[0078] In some embodiments, step S901 may include:

[0079] A plurality of first facial image samples of an object having a preset identity attribute and a preset age attribute are collected, and the plurality of first facial image samples are classified according to a plurality of preset facial style classification attributes to obtain facial image sample sets corresponding to respective facial styles.

[0080] Specifically, the objects with preset identity attributes and preset age attributes can be, for example, male and female celebrities in the age range of 15 to 30. Male and female celebrities are selected because the facial features of celebrities can basically represent the different definitions of beauty in different periods. Therefore, terminal 110 can collect multiple facial image samples of such objects with preset identity attributes and preset age attributes to obtain multiple first facial image samples. Terminal 110 can also further improve the breadth and accuracy of the data source by taking measures such as selecting and scoring the first facial image samples. After obtaining multiple first facial image samples, terminal 110 can classify these first facial image samples according to multiple preset facial style classification attributes, thereby forming facial image sample sets corresponding to various facial styles. For example, the preset facial style classification attributes can include facial style classification attributes such as different regions and different genders. That is, terminal 110 can classify the aforementioned multiple first facial image samples into different facial styles according to different regions and different genders to form corresponding facial image sample sets.

[0081] In some other embodiments, step S901 may include:

[0082] A plurality of second facial image samples imported by a user are obtained, and the plurality of second facial image samples are used as a facial image sample set corresponding to the custom facial style.

[0083] like Figure 3 As shown, terminal 110 can provide a user-defined style image import interface on first interface 310, allowing users to customize their desired facial style by importing images that match their desired facial deformation. Specifically, terminal 110 obtains multiple second facial image samples imported by the user and directly classifies the multiple second facial image samples into a facial image sample set corresponding to the customized facial style.

[0084] Step S902 : determining common facial features of a plurality of facial image samples for the facial image sample sets corresponding to the respective facial styles, and obtaining the common facial features corresponding to the respective facial styles.

[0085] This step mainly involves the terminal 110 analyzing the common facial features of multiple facial image samples contained in the facial image sample set corresponding to each facial style, and obtaining the common facial features corresponding to various facial styles. For example, the terminal 110 can obtain common facial features of different facial styles corresponding to different regions and different genders, and can also obtain common facial features corresponding to user-defined facial styles, etc., to provide users with rich and personalized facial deformation scheme options.

[0086] Furthermore, in some embodiments, Figure 10 As shown, determining the common facial features of the plurality of facial image samples in step S902 specifically includes:

[0087] Step S1001: Obtain facial feature samples of each facial image sample included in the facial image sample set.

[0088] In this step, after obtaining the facial image sample set, the terminal 110 can use the facial recognition algorithm to analyze and obtain the facial features (called facial feature samples) of each facial image sample in the facial image sample set, so as to facilitate the subsequent analysis of the common facial features of the facial image sample sets of each facial style.

[0089] Step S1002: input the facial feature samples of each facial image sample into a pre-built facial common feature recommendation model, and obtain the recommendation probability value of the facial image sample set output by the facial common feature recommendation model for the initial facial common feature.

[0090] In this step, the terminal 110 inputs the facial feature samples of each facial image sample into a pre-constructed facial common feature recommendation model, which is constructed based on the preset initial facial common features. The terminal 110 inputs the facial feature samples of each facial image sample into the model so that the facial common feature recommendation model obtains a recommendation probability value for the aforementioned preset initial facial common features based on the input facial feature samples. The recommendation probability value represents the degree of recommendation of the input facial feature samples for the preset initial facial common features. The terminal 110 obtains the recommendation probability value of the facial image sample set output by the model for the initial facial common features. The recommendation probability value can specifically indicate the probability of the aforementioned preset initial facial common features appearing in the facial feature samples of these facial image samples. The higher the probability of appearance, the higher the degree of recommendation of these facial image sample sets for the initial facial common features.

[0091] Among them, the facial common feature recommendation model can adopt models such as LR, FM, DNN, W&D, DeepFM, DIN, etc. Taking LR (i.e., Logistic Regression model) as an example, the terminal 110 can use the range of facial feature data currently recognized by the industry as the aforementioned preset initial facial common features, and generate training data with a time span of 5 years based on facial feature data from different regions and different genders, and use the Spark version of the LR algorithm to train a binary classification model for each facial style. If the imported training data meets the initial facial common features within a given time, it will be used as a positive example; otherwise, it will be used as a negative example. The trained model can be saved as a facial common feature recommendation model for predicting the probability of occurrence of the initial facial common features. Therefore, after the facial feature samples of each facial image sample are input into the pre-built facial common feature recommendation model, the recommendation probability value of the facial feature samples of each input facial image sample for the initial facial common features can be obtained based on the output result of the facial common feature recommendation model.

[0092] Step S1003: If the recommendation probability value is greater than or equal to the recommendation probability threshold, the initial facial common features are used as facial common features.

[0093] Specifically, if the recommendation probability value output by the facial common feature recommendation model is greater than or equal to the recommendation probability threshold, it means that the probability of the aforementioned preset initial facial common feature appearing in the facial feature samples of each input facial image sample exceeds a probability threshold of, for example, 85%. The terminal 110 can use the initial facial common feature as the facial common feature corresponding to the facial style. If the recommendation probability value is less than the recommendation probability threshold, the terminal 110 can further train and adjust the initial facial common feature until it can meet the requirements of the recommendation probability threshold condition.

[0094] The solution provided in this embodiment can use image recognition and machine learning / deep learning technologies in artificial intelligence to objectively, efficiently and accurately analyze the common facial features corresponding to more scientific and reasonable facial styles, avoid unscrupulous organizations from maliciously recommending facial deformation solutions, and help to objectively, scientifically and reasonably recommend facial deformation solutions to users when the users do not understand or are not sure about the expected effects.

[0095] The present application also provides an application scenario, which applies the above-mentioned facial deformation recommendation method based on image recognition. Specifically, the application scenario is to recommend facial plastic surgery solutions to users, and the solutions can be implemented on the terminal 110 based on interaction with the user, such as Figure 11 The flow chart of the recommended method is shown in FIG. Figure 3 、 5 8 describe the main process of this method, including:

[0096] The terminal 110 displays a first interface 310 for the user to select a facial style. After the user selects a facial style, the terminal 110 will display a facial template image 511 represented by the common facial feature K1 corresponding to the style on the second interface 510. The user can adjust and modify the position parameters of the common facial features by himself. The terminal 110 will provide a real-time preview of the modification effect during the user's parameter adjustment process. Then the user can click "Compare with me" on the second interface 510, and the terminal 110 enters the third interface 610 for the user to input the facial image to be deformed by importing a picture or taking a photo with the camera. Then the user can click "Start Comparison" on the third interface 610. The terminal 110 carefully compares the facial features of the facial image to be deformed and the common facial features of the aforementioned selected facial style to generate corresponding facial deformation recommendation information (or facial deformation recommendation plan), and provides a fourth interface 710 to display the facial deformation recommendation plan. The user can modify and adjust the data in the facial deformation recommendation plan on the fourth interface 710. If the user is not satisfied with the facial deformation recommendation plan and chooses not to manually adjust the plan, the terminal 110 returns to the first interface 310 for the user to reselect the facial style; if the user adjusts the facial deformation recommendation plan on the fourth interface 710 and clicks "Effect Preview", the terminal 110 provides a fifth interface 810, on which the facial deformation effect diagram is displayed.

[0097] The above scheme supports multiple ways of importing facial graphics to be deformed. Specifically, it can support multiple ways such as cameras and pictures to realize the collection of user's multi-angle facial information, and meet the needs of users collecting facial information in situations where it is inconvenient to turn on the camera. In the process of recommending facial deformation schemes, users can adjust the schemes by themselves. For example, after the terminal 110 displays the recommended facial deformation scheme, users can also modify the parameters of the scheme or even delete or add some adjustment items, providing users with more independent and customized capabilities. At the same time, it can also avoid the defects of traditional technologies in that facial deformation effects are all the same and lack personality. In addition, this method also allows users to view the expected effect of the current scheme in real time after adjusting the facial deformation scheme by themselves, which can support users to use more accurate facial deformation schemes to achieve the expected effect.

[0098] Furthermore, the above method can be based on Figure 12 The system architecture shown is implemented, specifically:

[0099] The data processing module is mainly responsible for the collection and processing of facial image sample data. During the sample data collection stage, the system can classify and collect data according to attributes such as different age groups, different regions, and different genders. Then, the facial features are analyzed through facial recognition algorithms, and these facial feature data are processed and analyzed. Among them, since a person's facial features will partially change at different ages, the data processing stage can mainly select male and female celebrities between the ages of 15 and 30. The main reason for selecting male and female celebrities is that the appearance of celebrities can basically represent the different definitions of beauty of people in different periods. The system may subsequently improve the breadth and accuracy of the data source by taking measures such as selecting and scoring facial image sample data. After collecting the corresponding facial image sample data, the data post-processing module organizes the facial features of different regions and genders into facial image sample sets belonging to different facial styles to facilitate further processing by the intelligent recommendation module.

[0100] For the intelligent recommendation module, its main content is model training and model prediction. Models such as LR, FM, DNN, W&D, DeepFM, DIN, etc. can be used. LR (Logistic Taking the logistic regression model as an example, the data extraction processing module generates a set of facial features of different regions and genders to generate training data with a time span of 5 years. The data range of beautiful facial features currently recognized in the industry can be used as the initial facial common features. Combined with the obtained training data, the Spark version of the LR algorithm is used to train a binary classification model for each facial style. If the imported training data meets this initial facial common feature within a given time, it is used as a positive example, otherwise it is used as a negative example. The trained model is saved and used to predict the probability of occurrence of the initial facial common feature. After the facial feature samples of each facial image sample are input into the facial common feature recommendation model, the recommended probability value output by the facial common feature recommendation model is obtained. If the recommended probability value output by the facial common feature recommendation model is greater than or equal to the recommended probability threshold, the initial facial common feature is used as the facial common feature corresponding to the facial style. If the threshold condition is not met, the initial facial common feature is further trained and adjusted until the constructed recommendation model can meet the requirements of the recommendation probability threshold condition.

[0101] By generating facial feature data for different regions and genders through the data processing module and conducting model training separately, recommendation models (or prediction models) for facial styles of different regions and genders can be generated. By doing so, common facial features (or recommended features) of facial styles of different regions and genders can be obtained, and the facial styles corresponding to these common facial features can be provided for users to choose.

[0102] For the solution generation module, combined with Figure 13Demonstrating its functional features, during the solution generation phase, users can choose from different facial styles pre-provided by the system, or user customization is supported. User customization is mainly achieved through the user importing images of the desired facial style or other methods that can be used to analyze and extract facial features. The system can analyze the user-imported photos and generate common facial features corresponding to the user-defined style. Then, the user imports his or her own facial information (corresponding to the facial image to be deformed) through a camera or photo. The system analyzes the facial features of the facial image to be deformed based on the facial recognition algorithm and compares them with the common facial features corresponding to the user-selected facial style or the user-defined style. Based on the main differences between the features, a facial deformation recommendation solution is generated, such as raising the nose bridge by 1mm, etc. At the same time, the generated facial deformation recommendation solution can support user adjustments and modifications, such as deleting some content, etc.

[0103] For the effect preview module, the effect preview module will generate the expected result of facial deformation based on the facial image to be deformed and the facial deformation recommendation scheme provided by the user. If the user modifies the facial deformation recommendation scheme by himself during the recommendation scheme generation stage, the system can also make real-time adjustments based on the user's modified facial deformation recommendation scheme.

[0104] Overall, the facial plastic surgery recommendation and effect preview system based on facial recognition algorithm and deep learning provided by this application can perform facial recognition algorithm analysis on a large number of images of celebrities from different regions and genders, train a recommendation model through a large amount of analysis data, and intelligently calculate facial plastic surgery templates with different facial styles for users to choose from. At the same time, the latest photos can be continuously imported to continuously improve the scientificity and reliability of the facial plastic surgery templates. After selecting facial plastic surgery templates with different facial styles, users can input their own multi-angle images through a camera or picture import. The system analyzes and compares and gives facial plastic surgery recommendations. At the same time, users can modify the facial plastic surgery recommendations and view the expected results. Its beneficial effects may include:

[0105] ① To provide users with more scientific facial plastic surgery plans, avoid malicious recommendations from some unscrupulous beauty institutions, and help recommend objective, scientific and reasonable facial deformation plans to users when users are not familiar with or confident about the expected results;

[0106] ② Provide users with a richer and more personalized selection of facial plastic surgery options. The system not only allows users to select facial styles based on different regions and genders, but also supports users to customize their desired styles by importing desired photos or other methods that can perform facial feature recognition and extraction.

[0107] ③ After the facial plastic surgery recommendation plan is generated, the system can also support users to modify the facial plastic surgery recommendation plan according to their personal wishes and view the expected results in real time, which can meet the user's more personalized needs.

[0108] It should be understood that, although the steps in the above flowchart are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the above flowchart may include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily to be performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0109] In one embodiment, Figure 14 As shown, a facial deformation recommendation device based on image recognition is provided. The device 1400 can be implemented as a software module or a hardware module, or a combination of the two to form a part of a computer device. The device specifically includes:

[0110] Style display module 1401, for displaying multiple facial styles;

[0111] Information acquisition module 1402 is configured to, in response to a style selection instruction for at least one of the plurality of facial styles, acquire facial deformation recommendation information corresponding to a facial feature comparison result obtained by comparing facial features corresponding to a facial image to be deformed provided by a user with common facial features corresponding to the selected facial style;

[0112] The deformation display module 1403 is used to display the facial deformation recommendation information and the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed.

[0113] In one embodiment, the information acquisition module 1402 is used to display a facial template image represented by the common facial features in response to the style selection instruction, and adaptively change the display of the facial template image when the common facial features are adjusted by the user; in response to a facial feature comparison instruction from the user, obtain facial deformation recommendation information corresponding to the facial feature comparison result; wherein the facial feature comparison result includes the comparison result of the facial common features not adjusted by the user and the facial features corresponding to the facial image to be deformed, or the comparison result of the facial common features adjusted by the user and the facial features corresponding to the facial image to be deformed.

[0114] In one embodiment, the information acquisition module 1402 is used to provide the common facial feature point position parameters corresponding to the common facial features on the facial template image; and adaptively change the display of the facial template image according to the user's adjustment of the common facial feature point position parameters.

[0115] In one embodiment, the deformation display module 1403 is used to obtain adjusted facial deformation recommendation information based on the user's adjustment of the facial deformation recommendation information; and in response to an effect preview instruction from the user, display the facial deformation effect obtained by applying the adjusted facial deformation recommendation information to the facial image to be deformed.

[0116] In one embodiment, the above-mentioned device 1400 also includes: a feature acquisition unit, used to obtain a facial image sample set corresponding to each facial style; the facial image sample set includes multiple facial image samples that meet preset recommendation conditions; for the facial image sample set corresponding to each facial style, determine the common facial features of the multiple facial image samples to obtain the common facial features corresponding to each facial style.

[0117] In one embodiment, a feature acquisition unit is configured to collect a plurality of first facial image samples of an object having a preset identity attribute and a preset age attribute; the preset recommendation condition includes the object having the preset identity attribute and the preset age attribute; and the plurality of first facial image samples are classified according to a plurality of preset facial style classification attributes to obtain facial image sample sets corresponding to the respective facial styles.

[0118] In one embodiment, the feature acquisition unit is configured to acquire a plurality of second facial image samples imported by a user; the preset recommendation condition includes the user importing; and the plurality of second facial image samples are used as a facial image sample set corresponding to the custom facial style.

[0119] In one embodiment, a feature acquisition unit is used to obtain facial feature samples of each facial image sample included in the facial image sample set; input the facial feature samples of each facial image sample into a pre-constructed facial common feature recommendation model, and obtain the recommendation probability value of the facial image sample set for the initial facial common feature output by the facial common feature recommendation model; the facial common feature recommendation model is constructed based on a preset initial facial common feature; if the recommendation probability value is greater than or equal to a recommendation probability threshold, the initial facial common feature is used as the facial common feature.

[0120] The specific limitations of the image recognition-based facial morphology recommendation device can be found in the limitations of the image recognition-based facial morphology recommendation method described above and will not be further elaborated here. Each module in the image recognition-based facial morphology recommendation device can be implemented in whole or in part via software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, allowing the processor to call and execute the corresponding operations of each module.

[0121] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 15 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a facial deformation recommendation method based on image recognition is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0122] Those skilled in the art will understand that Figure 15 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0123] In one embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0124] In one embodiment, a computer-readable storage medium is provided, storing a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0125] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps of each of the above-described method embodiments.

[0126] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0127] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions.

[0128] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0129] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A facial deformation recommendation method based on image recognition, characterized in that: The method comprises: Showcase a variety of facial styles; In response to a style selection instruction for at least one of the plurality of facial styles, obtaining facial deformation recommendation information corresponding to a facial feature comparison result obtained by comparing facial features corresponding to a facial image to be deformed provided by a user with common facial features corresponding to the selected facial style; The facial deformation recommendation information is displayed, and a facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed is displayed.

2. The method according to claim 1, characterized in that The step of obtaining facial deformation recommendation information corresponding to a facial feature comparison result in response to a style selection instruction for at least one of the multiple facial styles includes: In response to the style selection instruction, displaying a facial template image represented by the common facial features, and adaptively changing the display of the facial template image when the common facial features are adjusted by the user; In response to a facial feature comparison instruction from the user, obtaining facial deformation recommendation information corresponding to the facial feature comparison result; Among them, the facial feature comparison result includes the comparison result of the facial common features not adjusted by the user and the facial features corresponding to the facial image to be deformed, or the comparison result of the facial common features adjusted by the user and the facial features corresponding to the facial image to be deformed.

3. The method according to claim 2, characterized in that Adaptively changing the display of the facial template image when the common facial features are adjusted by the user includes: Providing the facial common feature point position parameters corresponding to the facial common feature on the facial template image; According to the user's adjustment of the position parameters of the common facial feature points, the display of the facial template image is adaptively changed.

4. The method according to claim 1, wherein The displaying of the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed includes: obtaining adjusted facial deformation recommendation information according to the user's adjustment of the facial deformation recommendation information; In response to an effect preview instruction from the user, a facial deformation effect obtained by applying the adjusted facial deformation recommendation information to the facial image to be deformed is displayed.

5. The method according to claim 1, wherein Before displaying a plurality of facial styles, the method further includes: Obtaining a facial image sample set corresponding to each facial style; the facial image sample set includes a plurality of facial image samples that meet a preset recommendation condition; For the facial image sample sets corresponding to the respective facial styles, common facial features possessed by the plurality of facial image samples are determined to obtain the common facial features corresponding to the respective facial styles.

6. The method according to claim 5, characterized in that The acquiring of facial image sample sets corresponding to the facial styles includes: collecting a plurality of first facial image samples of an object having a preset identity attribute and a preset age attribute; The preset recommendation conditions include the preset identity attribute and the preset age attribute; Classifying the plurality of first facial image samples according to a plurality of preset facial style classification attributes to obtain facial image sample sets corresponding to the respective facial styles; or, The acquiring of facial image sample sets corresponding to the facial styles includes: Acquire a plurality of second facial image samples imported by a user; the preset recommendation condition includes the second facial image samples imported by the user; The plurality of second facial image samples are used as a facial image sample set corresponding to the custom facial style.

7. The method according to claim 5 or 6, characterized in that Determining the common facial features of the plurality of facial image samples includes: Obtaining facial feature samples of each facial image sample included in the facial image sample set; Inputting the facial feature samples of each facial image sample into a pre-constructed facial common feature recommendation model, and obtaining a recommendation probability value of the facial image sample set for the initial facial common features output by the facial common feature recommendation model; the facial common feature recommendation model is constructed based on the preset initial facial common features; If the recommendation probability value is greater than or equal to the recommendation probability threshold, the initial facial common features are used as the facial common features.

8. A facial deformation recommendation device based on image recognition, characterized in that: The device comprises: Style display module, used to display multiple facial styles; an information acquisition module configured to, in response to a style selection instruction for at least one of the plurality of facial styles, acquire facial deformation recommendation information corresponding to a facial feature comparison result obtained by comparing facial features corresponding to a facial image to be deformed provided by a user with common facial features corresponding to the selected facial style; The deformation display module is used to display the facial deformation recommendation information and the facial deformation effect obtained by applying the facial deformation recommendation information to the facial image to be deformed.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. 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 method according to any one of claims 1 to 7 are implemented.

Images