Method and system for implementing makeup effect suggestion and storage medium

Abstract

The present invention provides a method for implementing makeup effect suggestions in a computer device, a system for implementing makeup effect suggestions, and a non-transitory computer-readable storage medium. The computer device is used to generate a digital image set describing a makeup effect and the makeup effect represents a makeup trend, and analyze the digital image set and extract target attributes. The computer device constructs a database of makeup suggestion entries and the database includes a digital image set and the extracted attributes, and receives a query request including a facial image of the user from a user. The computer device queries the database and obtains a first number of makeup suggestions. The computer device merges multiple makeup suggestions in the first number of makeup suggestions to generate a second number of makeup suggestions, displays at least a portion of the second number of makeup suggestions, and receives a selection from the user. The computer device executes a virtual application corresponding to the selected makeup effect.

Description

Technical Field

[0001] The present invention claims priority to and the benefit of U.S. Provisional Patent Application No. 62 / 875,581 filed on July 18, 2019, which is entitled “System and method for retrieving makeup recommendations based on facial features and makeup trends”, the entire contents of which are incorporated as part of the patent specification of the present invention for reference.

[0002] The present invention relates to, in particular to, a makeup effect suggestion system and method, and in particular to a system and method for suggesting makeup effects based on makeup trends and facial analysis. Background Art

[0003] Consumers invest a lot of money in makeup tools and accessories. Although online makeup tutorials are widely available, it can be challenging to achieve the same results as a professional makeup artist because of the wide selection of cosmetic products for different facial features. In addition, the makeup effects in such tutorials are not necessarily tailored to each person's specific facial features. Summary of the invention

[0004] One embodiment is a method for implementing makeup effect suggestions in a computer device, a computer device generates a set of digital images for describing makeup effects and the makeup effects represent makeup trends, analyzes the set of digital images, and extracts target attributes. The computer device constructs a database of makeup suggestion entries and the database includes the set of digital images and the extracted target attributes, and receives a query request including a facial image of the user from a user. The computer device queries the database and obtains a first number of makeup suggestions. The computer device merges multiple makeup suggestions in the first number of makeup suggestions to generate a second number of makeup suggestions, and displays at least a portion of the second number of makeup suggestions, and receives a selection from the user. The computer device executes a virtual application corresponding to the selected makeup effect.

[0005] Another embodiment is a system for implementing makeup effect suggestions, including a memory storing instructions and a processor coupled to the memory. The processor is configured by the instructions to generate a set of digital images describing makeup effects and the makeup effects represent makeup trends, analyze the set of digital images, and extract target attributes. The processor is further configured to construct a database of makeup suggestion entries and the database includes the set of digital images and the extracted target attributes, and receive a query request including a facial image of the user from a user. The processor is further configured to query the database and obtain a first number of makeup suggestions. The processor is further configured to merge multiple makeup suggestions in the first number of makeup suggestions to generate a second number of makeup suggestions, and display at least a portion of the second number of makeup suggestions, and receive a selection from the user. The processor is further configured to execute a virtual application corresponding to the selected makeup effect.

[0006] Another embodiment is a non-transitory computer-readable storage medium for storing instructions to be implemented by a computer device having a processor, wherein when the instructions are executed by the processor, the computer device generates a set of digital images depicting makeup effects and the makeup effects represent makeup trends, analyzes the set of digital images, and extracts target attributes. The processor is further configured to construct a database of makeup suggestion entries and the database includes the set of digital images and the extracted target attributes, and receives a query request including a facial image of the user from a user. The processor is further configured to query the database and obtain a first number of makeup suggestions. The processor is further configured to merge multiple makeup suggestions in the first number of makeup suggestions to generate a second number of makeup suggestions, and display at least a portion of the second number of makeup suggestions, and receive a selection from the user. The processor is further configured to execute a virtual application corresponding to the selected makeup effect.

[0007] By examining the following drawings and detailed description, other systems, methods, features and advantages of the present invention will be or become apparent to those skilled in the art. It is intended that all such additional systems, methods, features and advantages be included in the specification and scope of the present invention and protected by the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 is a block diagram of a computer device for suggesting makeup effects based on makeup trends and facial analysis according to various embodiments of the present invention.

[0009] Figure 2 yes Figure 1 Schematic diagram of a calculator device.

[0010] Figure 3 for Figure 1A high level flow chart of an example of the functionality of a portion of a computer device for suggesting makeup effects based on makeup trends and facial analysis.

[0011] Figure 4 Show Figure 1 The calculator device combines the makeup suggestions in the first number (N) of makeup suggestions to generate a second number (denoted by M) of makeup suggestions.

[0012] Figure 5 Show Figure 1 The calculator device filters the second number (M) of makeup suggestions to generate a third number (denoted by K) of makeup suggestions. DETAILED DESCRIPTION

[0013] Although there are many resources available to consumers, it can be challenging to achieve the same effects as makeup professionals because there are many types of makeup for different facial features. In addition, the makeup effects in such tutorials may not be suitable for everyone's specific facial features. The present invention discloses various embodiments for dynamically suggesting makeup effects based on the user's specific facial attributes and current makeup trends.

[0014] A description of a system for suggesting makeup effects based on makeup trends and facial analysis is now provided, followed by a discussion of the operation of the components within the system. Figure 1 1 is a block diagram of a computer device 102 in which the present invention may be implemented. The computer device 102 may be a computer device implemented by, for example, but not limited to, a smart phone, a tablet computer, a notebook computer, etc.

[0015] The makeup suggestion application 104 is executed on a processor of the computing device 102 and includes an image processor 106, a database manager 108, a matching module 110, and a makeup effect applicator 112. The image processor 106 is used to compile a collection of digital images depicting makeup effects that represent current makeup trends, and to analyze the collection of digital images and extract target attributes. In some embodiments, the image processor 106 includes a web crawler service that is used to extract digital images from different sources, including but not limited to social media websites and online photo albums, etc.

[0016] The image processor 106 analyzes the digital image set and extracts target attributes using a deep neural network model that includes multiple digital image samples corresponding to characteristics of different people. The target attributes may include, for example, high-dimensional vectors, such as d-dimensional vectors (x1, x2...xd), which are used to describe overall facial features, iconic facial features, size and shape attributes of facial features, makeup effects, and hairstyles, etc. The target attribute is represented as a d-dimensional vector x, where x is a sequence of d real numbers:

[0017]

[0018] The d-dimensional vector x is used to describe the characteristics of the overall facial features, iconic facial features, size and shape attributes of facial features, makeup effects, and hairstyle.

[0019] For some embodiments, the deep neural network model is pre-trained or initialized to distinguish faces corresponding to different characteristics of the individual. The image processor 106 then fine-tunes the deep neural network model by minimizing the distance between each sample in the deep neural network model and the center of all digital image samples associated with the same characteristic of the individual. The image processor 106 further fine-tunes the deep neural network model by maximizing the similarity between matching faces and simultaneously maximizing the distance between non-matching faces.

[0020] Those of ordinary skill in the art will appreciate that digital images may be encoded in any of a variety of formats, including, but not limited to, JPEG (Joint Photographic Experts Group) files, TIFF (Tagged Image File Format) files, PNG (Portable Network Graphics) files, GIF (Graphics Interchange Format) files, BMP (Bitmap) files, or many other digital formats. Alternatively, the digital image may be derived from a video still image encoded in a format including but not limited to Moving Picture Experts Group (MPEG)-1, MPEG-2, MPEG-4, H.264, 3rd Generation Partnership Project (3GPP), 3GPP-2, standard definition video (SD-Video), high definition video (HD-Video), digital versatile disc (DVD) multimedia, video compact disc (VCD) multimedia, high definition digital versatile disc (HD-DVD) multimedia, digital television video / high definition digital television (DTV / HDTV) multimedia, Audio Video Interleave (AVI), digital video (DV), QuickTime (QT) file, Windows Media Video (WMV), Advanced Systems Format (ASF), Real Media (RM), Flash Media (FLV), MPEG Audio Layer III (MP3), MPEG Audio Layer II (MP2), Waveform Audio Format (WAV), Windows Media Audio (WMA), 360 degree video, 3D scanned model, or many other digital formats.

[0021] The database manager 108 constructs a database 116 of makeup suggestions 120 based on the information extracted from each digital image 118 by the deep neural network model, and the database 116 includes a collection of digital images 118 and related makeup products. The database manager 108 further generates queries from users. Users upload their facial images to the database manager 108 to initiate queries. Once the user uploads an image of the user's face, the database manager 108 automatically performs face detection and extracts facial features from the uploaded image using the initialized deep neural network model.

[0022] The database manager 108 then queries the database and obtains a first number (denoted by N) of makeup suggestions 120. Each entry in the database 116 may also include a timestamp when the image processor 106 collects each digital image 118, and a corresponding descriptor may also be stored for each digital image 118. For some embodiments, the first number (N) of makeup suggestions 120 may be obtained from the database 116 using a hierarchical navigable small world (HNSW) algorithm, thereby identifying the top N entries in the database 116 that represent the most similar facial feature matches.

[0023] The matching module 110 is then used to merge the makeup suggestions 120 in the first number (N) of makeup suggestions 120 to generate a second number (denoted by M) of makeup suggestions. The matching module 110 merges the makeup suggestions 120 if the distance between the pairs of facial features of the corresponding digital images is less than a predefined threshold. Specifically, if the k-th facial feature (denoted by f_k) is similar to the j-th facial feature (denoted by f_j), the matching module 110 makes an assumption that the k-th and j-th facial features belong to the same feature. Based on the similarity between these facial features, the matching module 110 also assumes that the respective makeup effects of the k-th facial feature and the j-th facial feature are similar. Figure 4 The merging of makeup suggestions 120 from a first number (N) of makeup suggestions 120 to generate a second number (denoted by M) of makeup suggestions is shown.

[0024] The matching module 110 not only presents makeup suggestions 120 corresponding to similar images or makeup suggestions 120 corresponding to similar combinations of makeup effects, but also achieves higher diversity in the makeup suggestions 120 presented to the user. To this end, the matching module 110 recursively merges or groups similar pairs of makeup suggestions 120 with corresponding facial features, where the link distance (f_k and f_j) between the pairs of facial features is less than a predefined threshold T. The predefined threshold T is related to the boundary setting derived in the fine-tuning stage of the deep neural network model described previously. The boundary is defined as the distance from a given facial feature (feature point) to the decision boundary.

[0025] The matching module 110 merges the pairs of makeup suggestions 120 having the facial features (f_k and f_j) together and forms a new group (denoted by g) including the two makeup suggestions 120 having the facial features (f_k and f_j). A new makeup suggestion 120 having the facial features (denoted by f_g) is created and designated as a representative makeup suggestion 120 of the newly formed group g, wherein the new facial feature f_g of the representative makeup suggestion 120 is created by calculating the average value (average (f_k, f_j)) of the facial features in the group g. Then, the matching module 110 randomly selects a makeup suggestion 120 in the group g (f_k and f_j) and designates the randomly selected makeup suggestion 120 as the representative makeup suggestion 120 of the group g, thereby avoiding presenting similar or redundant makeup suggestions 120.

[0026] The user may adjust the predefined threshold (T) used to decide whether to merge facial features to adjust the degree of diversity of the makeup suggestions 120. For example, if the user sets the value of the predefined threshold T to zero, it will result in no merging or grouping of the makeup suggestions 120. This is likely to result in makeup suggestions 120 being associated with similar faces or even with faces corresponding to the same characteristics (i.e., the same person). On the other hand, if the value of the predefined threshold (T) is set too high, makeup suggestions 120 associated with dissimilar faces or with faces corresponding to different characteristics (i.e., different people) may be merged or grouped together. In this way, the user may adjust the value of the predefined threshold (T) as needed to achieve a desired degree of diversity in the makeup suggestions 120.

[0027] Once the above-mentioned merging or grouping operation is performed by the matching module 110, a second number (denoted by M) of makeup suggestions 120 is generated, wherein the second number (M) may be less than the first number (N) due to the merging of the makeup suggestions 120. For some embodiments, the second number (M) of makeup suggestions are presented to the user, wherein the makeup suggestions 120 may include, for example, reference information about each image associated with each makeup suggestion 120. Such reference information may include, for example, a related short article URL (URL), a timestamp of a specific short article, short article content (e.g., a rating), and the like.

[0028] The makeup suggestions 120 presented to the user may also include makeup analysis results that specifically indicate specific makeup products for specific facial features (e.g., makeup products for eyebrows, eye makeup, blush, lipstick). The makeup suggestions 120 presented to the user may also include information related to a makeup try-on kit to allow the user to actually apply the selected makeup product.

[0029] After the user selects one of the makeup suggestions 120, the makeup effect applicator 112 is used to perform a virtual application of the selected one or more makeup effects on the user's digital image. In the event that the user selects multiple makeup suggestions 120, the user can select the specific makeup effects to be applied and the order in which the makeup effects are applied. Once the virtual application of the selected makeup effects is complete, the user can capture a screenshot to share the user's digital image with others.

[0030] For some embodiments, the list of (M) makeup suggestions 120 is further narrowed down to increase the likelihood that the user will select one or more of the makeup suggestions 120 presented to the user, and the matching module 110 applies filtering to the (M) makeup suggestions 120 to generate a third number (denoted by K) of makeup suggestions. The filtering may be performed based on the time when the user queries the database 116, user preferences, and current makeup trend information, etc. For such embodiments, the third number (K) of makeup suggestions 120 is presented to the user, and a selection of the desired makeup suggestion 120 is received. Figure 5 Filtering (M) makeup suggestions 120 to generate a third number (denoted by K) of makeup suggestions is shown.

[0031] Figure 2 Shows Figure 1 1 is a schematic block diagram of a computer device 102 in FIG. The computer device 102 may be implemented in any of a variety of wired and / or wireless computer devices, such as a desktop computer, a portable computer, a dedicated server computer, a multi-processor computer device, a smart phone, a tablet computer, etc. Figure 2 As shown, the computer device 102 includes a memory 214 , a processing device 202 , a plurality of input / output interfaces 204 , a network interface 206 , a display 208 , a peripheral interface 211 , and a mass storage 226 , wherein each of these components are connected across a local data bus 210 .

[0032] Processing device 202 may include any of several custom or commercial processors associated with computer device 102, a central processing unit (CPU) or coprocessor, a semiconductor-based microprocessor (in the form of a microchip), a macroprocessor, one or more application-specific integrated circuits (ASICs), a plurality of appropriately configured digital logic gates, and other well-known electrical configurations that include discrete components that can coordinate the overall operation of the computer system either individually or in various combinations.

[0033] The memory 214 may include any combination of volatile storage components (e.g., random access memory (RAM, such as DRAM and SRAM, etc.)) and non-volatile storage components (e.g., ROM, hard disk drive, magnetic tape, CDROM). The memory 214 typically includes a native operating system 216, one or more native application programs, or an emulated system for various operating systems and / or emulated hardware platforms, an emulated operating system, etc. For example, these applications may include special-purpose software that may include Figure 1 Some or all of the components of the computer device 102 shown. According to such an embodiment, the components are stored in the memory 214 and executed by the processing device 202, so that the processing device 202 performs the operations / functions disclosed in the present invention. It will be understood by those skilled in the art that the memory 214 can and usually will include other components, which are omitted for the sake of brevity. For some embodiments, the components in the computer device 102 can be implemented by hardware and / or software.

[0034] The input / output interface 204 provides any number of interfaces for the input and output of data. For example, where the computing device 102 includes a personal computer, these components may be connected to one or more user input / output interfaces 204, which may include a keyboard or a mouse, such as a computer. Figure 2 Display 208 may include a computer monitor, a plasma screen for a personal computer (PC), a liquid crystal display (LCD) on a handheld device, a touch screen, or other display device.

[0035] In the present specification, a non-transitory computer readable medium stores a program used by or in conjunction with an instruction execution system, apparatus or device. More specific examples of computer readable media may include, but are not limited to: a portable computer disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM, EEPROM or flash memory) and a portable compact disk read-only memory (CDROM) (optical).

[0036] See also Figure 3 As shown, Figure 3 is based on Figure 1 Flowchart 300 of various embodiments of the computer device 102 performing makeup trend and facial analysis to suggest makeup effects. It can be understood that Figure 3 Flowchart 300 merely provides examples of different types of functional arrangements that may be used to implement the operations of the various components of computer device 102. Alternatively, Figure 3 Flowchart 300 of FIG. 1 may be viewed as describing an example of steps of a method implemented in computing device 102 according to one or more embodiments.

[0037] although Figure 3 The flowchart 300 of FIG. 300 shows a specific execution order, but it should be understood that the execution order may be different from the order described. For example, the execution order of two or more blocks may be pieced together relative to the order shown. Moreover, Figure 3 Two or more blocks shown in succession in the embodiment may be executed simultaneously or partially simultaneously. It should be understood that all such changes are within the scope of the present invention.

[0038] At block 310, the computing device 102 generates a set of digital images depicting makeup effects representative of current makeup trends, and analyzes the set of digital images and extracts target attributes. For some embodiments, the set of digital images is generated using a web crawler service, wherein the web crawler service extracts at least some digital images from social media websites.

[0039] At block 320, the computing device 102 analyzes the set of digital images and extracts target attributes using a deep neural network model that includes a plurality of samples corresponding to characteristics of different people. The target attributes may include, for example, high-dimensional vectors that describe overall facial features, iconic facial features, size and shape attributes of facial features, makeup effects, hairstyles, and the like.

[0040] For some embodiments, the deep neural network model is pre-trained or initialized to distinguish faces corresponding to different characteristics of the individual. The computing device 102 then fine-tunes the deep neural network model by minimizing the distance between each sample in the deep neural network model and the center of all digital image samples associated with the same characteristic of the individual. The computing device 102 also fine-tunes the deep neural network model by maximizing the similarity between matching faces while maximizing the distance between non-matching faces.

[0041] At block 330, computing device 102 constructs database 116 of entries ( Figure 1 ), the entry including the set of digital images 118 and the extracted attributes. For some embodiments, each entry in the database 116 may also include a timestamp when each digital image 118 was collected. A corresponding descriptor may also be stored for each digital image 118. At block 340, the computing device 102 receives a query generated by a user uploading an image of the user's face. At block 350, the computing device 102 queries the database and obtains a first number of makeup suggestions 120.

[0042] At block 360 , the computing device 102 incorporates the first quantity suggestion 120 ( Figure 1) to generate a second number of suggestions. For some embodiments, computer device 102 merges the makeup suggestions to generate the second number of makeup suggestions by grouping faces in database 116 that meet a similarity threshold, and for each group of faces, randomly selects a single face as a representative face. For some embodiments, each face in database 116 describes a corresponding makeup effect, and computer device 102 groups the faces in the database by grouping faces describing makeup effects that exhibit a similarity threshold into the same category and calculating an average value for each category to generate a representative value for each category. For some embodiments, each of the second number of makeup suggestions may include social media information and / or a suggested makeup product.

[0043] At block 370, the computing device 102 displays at least a portion of the second number of makeup suggestions and receives a user selection. At block 380, the computing device 102 executes a virtual application for a makeup effect corresponding to the selection. Figure 3 The processing is finished.

[0044] For some embodiments, the computer device 102 may be further configured to filter the second number of makeup suggestions to generate a third number of makeup suggestions, wherein the displayed portion of the second number of makeup suggestions includes the third number of makeup suggestions. For such embodiments, the computer device 102 filters the second number of makeup suggestions to generate the third number of makeup suggestions by prioritizing the second number of makeup suggestions based on a timestamp associated with the query request, user preferences, and current makeup trends.

[0045] It should be emphasized that the embodiments disclosed in the present invention described above are merely possible examples of implementation methods proposed for a clear understanding of the principles of the present invention. Many changes and modifications may be made to the above embodiments without departing substantially from the spirit and principles of the present invention. All these modifications and variations are intended to be included within the scope of the present invention and are protected by the appended claims.

Claims

1. A method for implementing makeup effect suggestions in a computer device, characterized in that: The method comprises: generating a set of digital images describing makeup effects, the makeup effects representing makeup trends; Analyze the digital image set and extract target attributes, wherein the target attributes include a high-dimensional vector; wherein the step of analyzing the digital image set and extracting the target attributes includes: Pre-train a deep neural network model to distinguish faces corresponding to different characteristics of individuals; Fine-tuning the deep neural network model by minimizing the distance between each sample and the center of all samples having a common characteristic; and Fine-tuning the deep neural network model by maximizing the similarity between matching faces and maximizing the distance between non-matching faces, and performing extraction of the target attribute using the deep neural network model; Constructing a database of makeup suggestion items, the database comprising the digital image set and the extracted target attributes; receiving a query request from a user including a facial image of the user; querying the database and obtaining a first number of makeup suggestions; Merging a plurality of makeup suggestions in the first number of makeup suggestions to generate a second number of makeup suggestions; wherein merging the makeup suggestions to generate the second number of makeup suggestions comprises: Grouping faces that meet a similarity threshold in the database; and For each group of faces, randomly select one face as a representative face; displaying at least a portion of the second number of makeup suggestions and receiving a selection from the user; and A virtual application corresponding to the selected makeup effect is executed.

2. The method according to claim 1, characterized in that The method further includes filtering the second number of makeup suggestions to generate a third number of makeup suggestions, wherein the at least a portion of the displayed second number of makeup suggestions includes the third number of makeup suggestions.

3. The method according to claim 2, characterized in that Filtering the second number of makeup suggestions to generate the third number of makeup suggestions includes: The second number of makeup suggestions are prioritized based on at least one of: a timestamp associated with the query request, a user preference, and a current makeup trend.

4. The method according to claim 1, characterized in that: Generating the collection of digital images is performed using a web crawler service, wherein the web crawler service extracts at least some of the digital images from social media websites.

5. The method according to claim 1, characterized in that: The high-dimensional vector includes at least one representation of: facial contour, hairstyle, hair color, head shape, size and shape of facial features, and makeup effects.

6. The method according to claim 1, characterized in that Each of said entries in said database further includes a timestamp of when each said digital image was collected and a descriptor.

7. The method according to claim 1, characterized in that Each of the faces in the database depicts a corresponding makeup effect.

8. The method according to claim 7, characterized in that Grouping the faces in the database comprises: grouping faces showing makeup effects with the threshold similarity into the same category; and The mean for each of the categories is calculated to produce a representative value for each of the categories.

9. The method according to claim 1, characterized in that: Each of the second number of makeup suggestions includes at least one of the following: Social media information; and Suggested makeup products.

10. A system for implementing makeup effect suggestions, characterized in that: The system comprises: a memory for storing instructions; a processor coupled to the memory, and the processor is configured by the instructions to at least: generating a set of digital images depicting makeup effects, the makeup effects representing makeup trends; Analyze the digital image set and extract target attributes, wherein the target attributes include a high-dimensional vector; wherein the step of analyzing the digital image set and extracting the target attributes includes: Pre-train a deep neural network model to distinguish faces corresponding to different characteristics of individuals; Fine-tuning the deep neural network model by minimizing the distance between each sample and the center of all samples having a common characteristic; and Fine-tuning the deep neural network model by maximizing the similarity between matching faces and maximizing the distance between non-matching faces, and performing extraction of the target attribute using the deep neural network model; Constructing a database of makeup suggestion items, the database comprising the digital image set and the extracted target attributes; receiving a query request from a user including a facial image of the user; querying the database and obtaining a first number of makeup suggestions; Merging a plurality of makeup suggestions in the first number of makeup suggestions to generate a second number of makeup suggestions; wherein merging the makeup suggestions to generate the second number of makeup suggestions comprises: Grouping faces that meet a similarity threshold in the database; and For each group of faces, randomly select one face as a representative face; displaying at least a portion of the second number of makeup suggestions and receiving a selection from the user; and A virtual application corresponding to the selected makeup effect is executed.

11. The system according to claim 10, characterized in that The processor is further configured to filter the second number of makeup suggestions to generate a third number of makeup suggestions, wherein the at least a portion of the displayed second number of makeup suggestions includes the third number of makeup suggestions.

12. The system according to claim 11, characterized in that The processor prioritizes the second number of makeup suggestions based on at least one of: a timestamp associated with the query request, a user preference, and a current makeup trend, to filter the second number of makeup suggestions to generate the third number of makeup suggestions.

13. The system according to claim 10, characterized in that The processor groups faces satisfying a similarity threshold in the database, and for each group of faces, randomly selects one of the faces as a representative face to merge the makeup suggestions to generate the second number of makeup suggestions.

14. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium is used to store instructions to be implemented by a computer device having a processor, wherein when the instructions are executed by the processor, the computer device at least: generating a set of digital images depicting makeup effects, the makeup effects representing makeup trends; Analyze the digital image set and extract target attributes, wherein the target attributes include a high-dimensional vector; wherein the step of analyzing the digital image set and extracting the target attributes includes: Pre-train a deep neural network model to distinguish faces corresponding to different characteristics of individuals; Fine-tuning the deep neural network model by minimizing the distance between each sample and the center of all samples having a common characteristic; and Fine-tuning the deep neural network model by maximizing the similarity between matching faces and maximizing the distance between non-matching faces, and performing extraction of the target attribute using the deep neural network model; Constructing a database of makeup suggestion items, the database comprising the digital image set and the extracted target attributes; receiving a query request from a user including a facial image of the user; querying the database and obtaining a first number of makeup suggestions; Merging a plurality of makeup suggestions in the first number of makeup suggestions to generate a second number of makeup suggestions; wherein merging the makeup suggestions to generate the second number of makeup suggestions comprises: Grouping faces that meet a similarity threshold in the database; and For each group of faces, randomly select one face as a representative face; displaying at least a portion of the second number of makeup suggestions and receiving a selection from the user; and A virtual application corresponding to the selected makeup effect is executed.

15. The non-transitory computer readable storage medium of claim 14, wherein: The processor is further configured to filter the second number of makeup suggestions to generate a third number of makeup suggestions, wherein the at least a portion of the displayed second number of makeup suggestions includes the third number of makeup suggestions.

16. The non-transitory computer-readable storage medium of claim 15, wherein: The processor prioritizes the second number of makeup suggestions based on at least one of: a timestamp associated with the query request, a user preference, and a current makeup trend, to filter the second number of makeup suggestions to generate the third number of makeup suggestions.

Images