Commodity recommendation method and device, electronic equipment, storage medium and program product

Abstract

The invention provides a commodity recommendation method and device, electronic equipment, a computer readable storage medium and a computer program product, and relates to the technical field of computers. The method comprises the following steps: determining a to-be-recommended commodity type from a commodity recommendation request; extracting skin features corresponding to the to-be-recommended commodity type from the face image of the target user; and pushing commodity information of the to-be-recommended commodity type to the target user based on the skin features corresponding to the to-be-recommended commodity type. According to the method, a commodity recommendation mode which not only adapts to the commodity recommendation requirements of the user, but also considers the skin characteristics of the user can be automatically realized, and the commodity recommendation accuracy is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a product recommendation method, a product recommendation device, an electronic device, a computer-readable storage medium, and a computer program product. Background Art

[0002] Most existing product recommendation solutions make recommendations based on the user's historical purchase and browsing history, and users basically make purchases based on their personal preferences. As a result, existing product recommendation solutions require the acquisition of a large amount of historical data generated by users when purchasing and / or browsing products. On the one hand, data collection is difficult and the quality of the collected data is difficult to guarantee. On the other hand, the recommendation method based on user historical data is greatly affected by the user's personal aesthetic factors, and it is easy for the recommended products to be inappropriate for the user himself, that is, the recommendation results are inaccurate, thereby affecting the user's product usage experience. Summary of the invention

[0003] The embodiments of the present disclosure propose a product recommendation method, a product recommendation device, an electronic device, a computer-readable storage medium, and a computer program product, which can automatically implement a product recommendation method that is compatible with the user's product recommendation needs and takes into account the user's skin characteristics, thereby improving the accuracy of product recommendations.

[0004] In a first aspect, an embodiment of the present disclosure proposes a product recommendation method, including: determining the type of product to be recommended from a product recommendation request; extracting skin features corresponding to the type of product to be recommended from a face image of a target user; and pushing product information of the type of product to be recommended to the target user based on the skin features corresponding to the type of product to be recommended.

[0005] In a second aspect, the disclosed embodiment proposes a product recommendation device, comprising: a determination module, an extraction module and a push module. The determination module is configured to determine the type of product to be recommended from a product recommendation request; the extraction module is configured to extract skin features corresponding to the type of product to be recommended from a face image of a target user; and the push module is configured to push product information of the type of product to be recommended to the target user based on the skin features corresponding to the type of product to be recommended.

[0006] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can implement the product recommendation method described in any implementation method in the first aspect when executing.

[0007] In a fourth aspect, an embodiment of the present disclosure provides a non-transitory computer-readable storage medium storing computer instructions, which are used to enable a computer to implement the product recommendation method described in any implementation method of the first aspect when executed.

[0008] In a fifth aspect, an embodiment of the present disclosure provides a computer program product including a computer program, which, when executed by a processor, can implement the commodity recommendation method described in any implementation manner in the first aspect.

[0009] The product recommendation solution provided by the embodiment of the present disclosure needs to determine the type of product to be recommended from the product recommendation request of the user, and extract the skin features of the user that are compatible with the type of product to be recommended based on the user's face image, and then push the product information of the type of product to be recommended to the user based on the extracted skin features. In this way, a product recommendation method that is compatible with the user's product recommendation needs and takes into account the user's skin features can be automatically implemented, which is not limited to the user's personal aesthetics or preferences, but can meet the user's needs, thereby improving the accuracy of product recommendations.

[0010] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Other features, objects and advantages of the present disclosure will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0012] Figure 1 is an exemplary system architecture in which the present disclosure may be applied;

[0013] Figure 2 A flowchart of a product recommendation method provided by an embodiment of the present disclosure;

[0014] Figure 3 A flowchart of another product recommendation method provided by an embodiment of the present disclosure;

[0015] Figure 4 A structural block diagram of a commodity recommendation device provided by an embodiment of the present disclosure;

[0016] Figure 5 A schematic diagram of the structure of an electronic device suitable for executing a product recommendation method provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0017] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description. It should be noted that the embodiments in the present disclosure and the features in the embodiments may be combined with each other without conflict.

[0018] It should be pointed out that in the technical solution of the present invention, the collection, acquisition, storage, processing, transmission, provision, disclosure and application of user personal information (such as facial images, etc.) are all carried out with the user's knowledge and explicit authorization, comply with the relevant laws and regulations, and do not violate public order and good morals.

[0019] Figure 1 An exemplary system architecture 100 is shown to which embodiments of the product recommendation method, apparatus, electronic device, and computer-readable storage medium of the present disclosure can be applied.

[0020] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0021] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various applications for realizing information communication between the terminal devices 101, 102, 103 and server 105 can be installed, such as product recommendation applications, shopping applications, instant messaging applications, etc.

[0022] Terminal devices 101, 102, 103 and server 105 can be hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices with display screens, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, etc.; when terminal devices 101, 102, 103 are software, they can be installed in the electronic devices listed above, which can be implemented as multiple software or software modules, or as a single software or software module, and no specific limitation is made here. When server 105 is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server; when the server is software, it can be implemented as multiple software or software modules, or as a single software or software module, and no specific limitation is made here.

[0023] The server 105 can provide various services through various built-in applications. For example, the server 105 can provide a product recommendation application based on the user's face image. When running the product recommendation application, the server 105 can achieve the following effects: it is necessary to determine the type of product to be recommended from the user's product recommendation request, and extract the user's skin features that are compatible with the type of product to be recommended based on the user's face image, and then push the product information of the type of product to be recommended to the user based on the extracted skin features. In this way, a product recommendation method that is compatible with the user's product recommendation needs and takes into account the user's skin features can be automatically implemented, which is not limited to the user's personal aesthetics or preferences, but can meet the user's needs, thereby improving the accuracy of product recommendations.

[0024] It should be noted that the server 105 can obtain the product information through the network 104 or store it locally in advance in various ways. Therefore, when the server 105 detects that the data has been stored locally (for example, various types of product information have been pre-stored), it can choose to directly obtain the data from the local.

[0025] Since extracting skin features, pushing product information, etc. require more computing resources and stronger computing power, the product recommendation method provided in the subsequent embodiments of the present disclosure is generally performed by the server 105 with stronger computing power and more computing resources. Accordingly, the product recommendation device is generally also set in the server 105. However, it should also be pointed out that when the terminal devices 101, 102, and 103 also have computing power and computing resources that meet the requirements, the terminal devices 101, 102, and 103 can also complete the above-mentioned operations that are handed over to the server 105 through the product recommendation application installed thereon, and then output the same result as the server 105. Especially in the case where there are multiple terminal devices with different computing powers at the same time, but when the product recommendation application determines that the terminal device where it is located has stronger computing power and more remaining computing resources, it can let the terminal device perform the above-mentioned operations, thereby appropriately reducing the computing pressure of the server 105. Accordingly, the product recommendation device can also be set in the terminal devices 101, 102, and 103. In this case, the exemplary system architecture 100 may also not include the server 105 and the network 104.

[0026] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0027] Please refer to Figure 2 , Figure 2 This is a flowchart of a product recommendation method provided by an embodiment of the present disclosure, wherein process 200 includes the following steps:

[0028] Step 201: Determine the type of product to be recommended from the product recommendation request.

[0029] This step is intended to be performed by the execution subject of the product recommendation method (e.g. Figure 1 The server 105 shown in the figure can receive the user's Figure 1 The terminal devices 101, 102, 103 shown in the figure) send a product recommendation request, and in response to receiving the product recommendation request, determine the type of product to be recommended from the product recommendation request.

[0030] In some optional implementations of the embodiments of the present disclosure, the above-mentioned product recommendation request can be used to indicate information related to the product (or item) that the user wants to know or buy. For example, the product recommendation request can carry the product name of the product to be recommended and / or the type of the product to be recommended that the user needs, so that the type of product to be recommended can be indirectly determined based on the product name or the type of product to be recommended can be directly determined. In some application scenarios, the user can log in to the system through the user interface of a website or platform with a product recommendation function (for example, the login can be verified by a user name and login password), and enter or select index keywords associated with the product to be recommended (such as product name or product type index keywords) to trigger the sending of the product recommendation request.

[0031] In some optional implementations of the disclosed embodiments, the above-mentioned types of goods to be recommended include at least one of the following: skin care products and cosmetics. Among them, skin care products may include but are not limited to: cleansing skin care products, such as facial cleansers, makeup remover oils, etc.; skin care or basic skin care products, such as toners, lotions, essences, face creams, eye creams, facial masks, sunscreens, etc. Cosmetics (also referred to as color cosmetics) include but are not limited to: lip cosmetics, such as lipsticks, lip glazes, lipsticks, lip muds, etc.; eye cosmetics, such as eyebrow pencils, eyeliners, eye shadows, mascara, false eyelashes, etc.; cheek area cosmetics, such as blush, highlights, etc.; whole face or full face cosmetics, such as foundation, primer, loose powder, air cushion, etc.

[0032] In some optional implementations of the embodiments of the present disclosure, the above-mentioned types of goods to be recommended may also include facial accessories, such as forehead pendants, facial accessories, face chains, etc.

[0033] Step 202: extracting skin features corresponding to the type of product to be recommended from the face image of the target user.

[0034] Based on step 201, this step is intended to be performed by the above-mentioned execution subject (for example Figure 1 The server 105 shown in the figure extracts skin features corresponding to the type of product to be recommended from the face image of the target user. Continuing with the above application scenario as an example, after logging into the system through the user interface of a website or platform with a product recommendation function, the user can upload a photo or picture including the user's face image, and the face image of the target user can be obtained from the photo or picture.

[0035] In some optional implementations of the embodiments of the present disclosure, a correspondence between different product types and different skin features can be pre-established, wherein the above-mentioned skin features include but are not limited to skin color features, texture features, pore features, stratum corneum features, scar features, etc., and further, at least one of the skin color features, texture features, pore features, stratum corneum features and scar features can be further refined and distinguished, so that corresponding skin care products and / or cosmetics can be matched according to different skin features, which is conducive to improving the accuracy of product recommendations; wherein the above-mentioned correspondence can be set according to specific application scenarios, actual needs, etc.

[0036] In some optional implementations of the embodiments of the present disclosure, the facial image of the target user can be accurately detected and extracted from the received photos or pictures based on a pre-trained face detection model, and then the skin features corresponding to the type of product to be recommended can be extracted based on the accurate locking of the facial image.

[0037] In some optional implementations of the embodiments of the present disclosure, the pre-trained face detection model can be trained through the following process:

[0038] Process 1): Obtain training sample data, wherein the training sample data includes a sample face image and annotation information of the sample face image, wherein the annotation information of the sample face image is used to indicate the facial key points contained in the corresponding sample face image, wherein the facial key points can be pre-set prominent feature points on the face, such as left and right eyeballs, nose tip, and mouth corners. In some examples, the training sample data can be selected from a face recognition benchmark dataset (WIDER FACE).

[0039] Process 2): Input the above sample face images into the face detection model to be trained to perform model training.

[0040] Among them, the face detection model to be trained can be a face detection model based on the RetinaFace network framework. Of course, other face detection frameworks can also be used to train the face detection model, such as a multi-task convolutional neural network (MTCNN) or Dlib (a C++ open source toolkit containing machine learning algorithms). Here, the face detection model training using the RetinaFace network framework is used as an example for explanation. For the input sample face image to be detected, the RetinaFace network first enters the backbone network, and the backbone feature network extraction can be performed using the MobileNet network or the ResNet50 network model. The basic features of the image can be obtained through the backbone network. Then, the basic features output by the backbone network are enhanced, and the feature pyramid (FPN) and SSH (Single Stage Headless) models can be used to enhance the feature extraction. Through the enhancement of features, the RetinaFace network can more clearly describe the facial features of the face. After feature extraction, we can use ClassHead, BoxHead, and LandmarkHead networks to obtain prediction results from the extracted features, and decode the prediction results through the Pytorch framework. Finally, we remove duplicate detection values ​​through non-maximum suppression (NMS) to obtain the final prediction results. The prediction results can be used to determine whether the current region of interest contains the facial shape of the face, and further obtain the key feature points of the face.

[0041] Process 3): Compare the prediction results with the labeled information and the information inferred from the labeled information, adjust the parameters, predict again, compare again, and repeat the process until the recognition accuracy reaches, for example, 89% or more. Then, a face detection model that can provide reasonable output can be trained.

[0042] In some other optional implementations of the embodiments of the present disclosure, in addition to the above-mentioned method of training a new face detection model, in order to reduce training time and data requirements, transfer learning can also be implemented to fine-tune the pre-trained face detection model to be migrated, such as the deep learning model VGGFace or FaceNet (for example, by adjusting (for example, increasing) the weights of some sample face images used to train the face detection model, or by adjusting (for example, increasing) the annotation information of the sample face images used to train the face detection model) to obtain a face detection model suitable for the embodiments of the present disclosure.

[0043] In some optional implementations of the disclosed embodiments, a preset feature detection method may be used to extract skin features corresponding to the type of product to be recommended from the face image of the target user. In this way, skin features that match the user's product recommendation needs may be accurately extracted.

[0044] In some optional implementations of the embodiments of the present disclosure, the above-mentioned preset feature detection method can be a 68 feature point detection method or a 468 feature point detection method, which can mark the feature points of the face such as the nose, eyes, eyebrows and facial contour one by one to represent the relevant features of the face, and can output the corresponding feature point numbers. In this way, the pixel position of the feature points of each important component of the face can be obtained, which can be used to detect the facial feature points of the face after face detection, and then accurately extract the skin features corresponding to the corresponding type of product to be recommended.

[0045] It should be noted that in other optional implementation methods of the embodiments of the present disclosure, other detection methods besides the above-mentioned 68 feature point detection method or 468 feature point detection method can also be used to extract skin features corresponding to the above-mentioned product types to be recommended, such as using a pre-trained skin feature extraction model, which is not specifically limited here.

[0046] Step 203: Pushing product information of the product type to be recommended to the target user based on the skin features corresponding to the product type to be recommended.

[0047] Based on step 202, this step is intended to be performed by the above-mentioned execution subject (for example Figure 1 The server 105 shown in the figure accurately pushes the product information of the product type to be recommended to the target user based on the skin features corresponding to the product type to be recommended extracted as mentioned above.

[0048] In some optional implementations of the disclosed embodiments, the above-mentioned commodity information may include but is not limited to at least one of the following: commodity brand information, commodity series information, commodity color information, commodity texture information, commodity name information, commodity nickname information, and commodity price information. Taking the above-mentioned application scenario as an example, the user can log in to the system based on the user interface of a website or platform with a commodity recommendation function, and after uploading a photo or picture including the user's face image to extract the skin features of the commodity type to be recommended, the above-mentioned commodity information pushed to the target user based on the skin features of the commodity type to be recommended can be further displayed through the corresponding display page of the system. Among them, the above-mentioned commodity information can be obtained from the relational database MySQL, or can also be obtained from a higher-performance database PostgreSQL (a free object-relational database) or a non-relational database (Not only SQL, NoSQL, such as MongoDB (a database based on distributed file storage)) to better handle large amounts of data and high concurrent requests.

[0049] The commodity recommendation method provided by the embodiment of the present disclosure needs to determine the commodity type to be recommended from the commodity recommendation request of the user, and extract the user's skin features that are compatible with the commodity type to be recommended based on the user's face image, and then push the commodity information of the commodity type to be recommended to the user based on the extracted skin features. In this way, a commodity recommendation method that is compatible with the user's commodity recommendation needs and takes into account the user's skin features can be automatically implemented, which is not limited to the user's personal aesthetics or preferences, but can meet the user's needs, thereby improving the accuracy of commodity recommendations.

[0050] Please refer to Figure 3 , Figure 3 This is a flowchart of another product recommendation method provided by an embodiment of the present disclosure. In the process 300 of the product recommendation method, the process 300 includes the following steps:

[0051] Step 301: Determine the type of product to be recommended from the product recommendation request.

[0052] Step 302: extracting skin features corresponding to the type of product to be recommended from the face image of the target user.

[0053] The above steps 301-302 are similar to Figure 2 The steps 201-202 shown in the figure are consistent. For the same contents, please refer to the corresponding parts of the previous embodiment and will not be repeated here. Further, in the case where the skin features corresponding to the above-mentioned product types to be recommended include skin color features, step 302 can also be specifically described as: extracting the skin color features corresponding to the product types to be recommended from the face image of the target user. Figure 2The step 203 shown can be implemented specifically through the following steps:

[0054] Step 303: Determine a first skin color feature vector based on the skin color feature in the skin features corresponding to the type of product to be recommended.

[0055] Step 304: Based on the skin color similarity between the first skin color feature vector and the second skin color feature vector of the candidate user, determine similar users of the target user from the candidate users. In this embodiment, the similar users of the target user have a high similarity with the target user at least in skin color. Further, in response to the number of candidate users being multiple, determining similar users of the target user from the candidate users may specifically include: sorting the multiple candidate users according to the calculated skin color similarity, and determining the candidate user who ranks first and has the highest similarity with the target user as a similar user of the target user. Further, in response to the number of candidate users who have the highest similarity with the target user, candidate users who are the same or similar to the target user in one or more of gender, age group, and region may be determined as similar users of the target user.

[0056] Step 305: Pushing product information to the target user based on the first historical product usage information of similar users; wherein the first historical product usage information corresponds to the type of product to be recommended.

[0057] In this embodiment, corresponding product information recommendation can be performed based on the skin color feature in the skin features of the target user. Specifically, the similarity between the skin colors of the user's face can be determined based on the skin color feature vectors corresponding to the skin color features of the target user and each candidate user. By evaluating the skin color similarity, similar users of the target user can be accurately screened from the candidate users from the perspective of similar skin color, thereby obtaining a similar relationship between users. Then, based on the similar relationship, the item information that is most likely to be pushed to the target user is selected from the behaviors of similar users of the target user. In other words, the accurate product information can be pushed to the target user based on the historical product usage data of the similar user corresponding to the above-mentioned product type to be recommended, that is, the first historical product usage information. In this way, through user-based collaborative filtering, user data, including product information used or purchased by users, is collected and sorted, and the similarity of skin color between users is further considered, and the relevant information of the corresponding type of product is recommended to the target user through the historical product usage data of similar users.

[0058] In this embodiment, product information is recommended by obtaining the first historical product usage information corresponding to the type of product to be recommended of similar users of the target user without obtaining the full amount of historical product usage data of the similar users. This can reduce the difficulty of data collection, and by controlling the amount of data collection, the quality of the collected data can be guaranteed to a certain extent, which can further improve the efficiency and accuracy of product recommendations.

[0059] In some optional implementations of the embodiments of the present disclosure, the skin color similarity between the above-mentioned first skin color feature vector and the second skin color feature vector of the candidate user can be characterized by calculating the cosine similarity between the skin color feature vectors, or by calculating the Pearson correlation between the skin color feature vectors, thereby weighing the skin color similarity between users.

[0060] In some optional implementations of the embodiments of the present disclosure, the first historical product usage information of the above-mentioned similar users may include but is not limited to at least one of the following information of the products used or purchased: product brand information, product series information, product color information, product texture information, product name information, product information, and product price information.

[0061] In some optional implementations of the embodiments of the present disclosure, the above-mentioned step 303 can be further specifically executed as the following contents, including: converting the first color space data of the skin color characteristics corresponding to the type of product to be recommended into second color space data; clustering the second color space data to obtain a first skin color feature vector.

[0062] In this embodiment, the skin color feature corresponding to the type of product to be recommended can be converted into color space data, that is, the original color space data with a large amount of data in the extracted skin color feature (that is, the first color space data, such as RGB color space data, where R, G, and B represent red, green, and blue, respectively, and the three channels of R, G, and B are used as the X, Y, and Z axes in the Cartesian coordinate system to obtain a spatial description of the color) is converted into the second color space data that can better characterize the user's skin color features (for example, HSV (hue or hue, saturation or color purity Saturation) tion, brightness Value) color space data, where hue H is measured by angle, with a value range of 0° to 360°, saturation S with a value range of 0.0 to 1.0, and brightness V with a value range of 0.0 (black) to 1.0 (white). The HSV color space model can also be called a hexagonal pyramid model; it can also be HSL color space data (also known as ), which is similar to HSV, where L represents lightness, and can also be called HSB, where B represents brightness; etc.), remove redundant information in the first color space data and improve the accuracy of skin color representation. In order to further reduce the amount of calculation and improve the accuracy of the skin color feature vector that represents the user's skin color characteristics, the converted second color space data can be clustered to cluster the feature information that can represent the user's skin color characteristics into a multi-dimensional skin color feature vector.

[0063] In some optional implementations of the embodiments of the present disclosure, a conversion scheme matching specific color space data may be used to implement conversion from first color space data to second color space data.

[0064] In some optional implementations of the embodiments of the present disclosure, a preset clustering algorithm may be used to perform clustering processing on the second color space data to obtain the first skin color feature vector. The preset clustering algorithm may include a K-means clustering algorithm (also called a K-Means clustering algorithm), a mean shift clustering algorithm, an expectation maximization (EM) clustering using a Gaussian mixture model (GMM), and the like.

[0065] Furthermore, taking the use of the above-mentioned K-means clustering algorithm to perform clustering processing on the second color space data to obtain the above-mentioned first skin color feature vector as an example for explanation, k centroids or center points for clustering are initialized and selected from the n samples of the second color space data, and the distances between other samples and each centroid (such as Euclidean distance) are calculated, so as to classify each sample into a category where the nearest centroid is located based on the principle of closest distance, re-determine the centroid, and repeat the above-mentioned steps of calculating the distance and performing clustering based on the distance until the results of the two most recent clusterings do not change, and finally obtain k classifications as the clustering results output, and then the corresponding color space data can be represented as a k-dimensional feature vector; wherein n and k are both positive integers.

[0066] It should be noted that the second skin color feature vector of each candidate user may be determined in the same manner as the first skin color feature vector, which will not be described in detail herein.

[0067] In some optional implementations of the embodiments of the present disclosure, the above step 305 can be further specifically executed as the following contents, including: determining a first product usage feature vector of a similar user based on the first historical product usage information; determining a second product usage feature vector of the target user based on the second historical product usage information of the target user; wherein the second historical product usage information corresponds to the type of product to be recommended; and determining the product information to be pushed to the target user from the first historical product usage information based on the similarity between the first product usage feature vector and the second product usage feature vector.

[0068] In this embodiment, after similar users of the target user are determined based on the similarity of skin color between the target user and the candidate user, the product recommendation to the target user can be further completed based on the similarity between the products. That is, the product usage feature vectors of the target user and the similar user (i.e., the first product usage feature vector and the second product usage feature vector) can be determined based on the historical product usage information of the target user and the similar user corresponding to the type of product to be recommended, and then the similarity between the two product usage feature vectors can be determined to evaluate the similarity of different users in product usage habits or preferences, so that the relevant information of the products of similar users with high similarity to the products used or purchased by the target user can be pushed to the target user to achieve personalized recommendation. In some implementations of the embodiments of the present disclosure, the product usage feature vector can be determined and the similarity can be evaluated based on historical product brand information, historical product texture information, and historical product price information.

[0069] In this embodiment, product information is recommended by obtaining the historical product usage information of the target user and its similar users corresponding to the type of product to be recommended, without obtaining the full amount of historical product usage data of the target user and its similar users. This can reduce the difficulty of data collection, and by controlling the amount of data collection, the quality of the collected data can be guaranteed to a certain extent, which can further improve the efficiency and accuracy of product recommendations.

[0070] It should be noted that the method for determining any of the above-mentioned product usage feature vectors and the method for calculating the similarity between product usage feature vectors can refer to the above-mentioned related schemes for determining skin color feature vectors and the scheme for calculating the similarity between skin color feature vectors, which will not be repeated here.

[0071] It should be noted that, for the above-mentioned skin features including other situations in addition to the above-mentioned skin color features, such as replacing the skin color features with at least one of the texture features, pore features, stratum corneum features, and scar features, or on the basis of the skin color features, further including at least one of the texture features, pore features, stratum corneum features, and scar features, the corresponding feature vector can be determined by a scheme similar to the above-mentioned implementation method of determining the skin color feature vector, and further based on the corresponding feature vector, the similarity between the target user and the candidate users is determined, and similar users of the target user are screened out from the candidate users. The specific process is not repeated here. In response to using multiple of the above-mentioned features for similarity determination to screen similar users, different features can be used in sequence to gradually narrow the scope for screening, or similar users of the target user can be determined through a one-time screening at the same time.

[0072] In the product recommendation method provided by any of the above-mentioned disclosed embodiments, the above-mentioned step 202 or step 302 can be further specifically implemented as the following contents, including: based on the association relationship between the pre-set product type and the facial area of ​​the human face, determining the target facial area corresponding to the type of product to be recommended from the facial image of the target user; based on the facial features of the detected target facial area, extracting the skin features corresponding to the type of product to be recommended.

[0073] In some optional implementations of the embodiments of the present disclosure, the above-mentioned pre-established correspondence (association) relationship between different product types and different skin features can be further specified as: pre-establishing an association relationship between different product types and skin features of different facial areas of the face, thereby establishing an association relationship between the product type, the facial area of ​​the face and the corresponding skin features. In this embodiment, when it is necessary to extract skin features corresponding to the product type to be recommended based on the facial image of the target user, the target facial area associated with the product type to be recommended can be locked based on the pre-set association relationship between the product type and the facial area of ​​the face, and then the skin features corresponding to the product type to be recommended can be extracted based on the detected facial features or facial feature points of the target facial area. In this way, skin features that are highly compatible with the user's product recommendation needs can be accurately extracted, thereby improving the accuracy of product recommendations.

[0074] In some optional implementations of the embodiments of the present disclosure, in the process of extracting skin features corresponding to the above-mentioned type of recommended goods from the facial image of the above-mentioned target user using the above-mentioned preset feature detection method (such as a 68-feature point detection method or a 468-feature point detection method), each facial feature point or facial feature of the face determined based on the preset feature detection method has its own number, and the components of the number may at least include "facial area number-feature point sequence number within the area". The feature data of the corresponding facial feature point or facial feature can be indexed by the number, which can be used to calculate the similarity of the corresponding features. Among them, the facial area of ​​the human face may include the cheek area, the mandibular area, the forehead area, and the area around the eyes. For example, the number 1-001 can represent the facial feature point or facial feature No. 001 detected in the forehead area. Then, in some optional implementation methods of the embodiments of the present disclosure, the facial feature points or facial features of the target facial area can be determined in the following way: detect a preset number of facial feature points or facial features of the face image; wherein the preset number of facial feature points or facial features are numbered according to the facial area of ​​the face image; extract the number of the target facial area corresponding to the type of product to be recommended from the number of the preset number of facial feature points; determine the facial feature points or facial features indicated by the number located in the target facial area (for example, by connecting the numbers to circle the target facial area to remove interference of other parts in the face image with the determination of the skin features of the target facial area) as the facial feature points or facial features corresponding to the type of product to be recommended.

[0075] In some optional implementations of the embodiments of the present disclosure, in response to the type of product to be recommended including cosmetics, the pre-established association relationship between the above-mentioned different product types and the skin features of different facial areas of the human face may include but is not limited to: in response to the type of product to be recommended being specifically lip cosmetics (such as lipstick, lip gloss, lip balm, lip mud, etc.), it can be associated with the skin features of the cheek area and / or jaw area of ​​the human face, and further, the skin features may include skin color features; in response to the type of product to be recommended being specifically eye cosmetics (such as eyebrow pencil, eyeliner, eye shadow, mascara, false eyelashes, etc.), it can be associated with the forehead area and / or the area around the eyes of the human face, and further, the skin features may include skin color features and / or texture features; in response to the type of product to be recommended being specifically whole face or full face cosmetics (such as liquid foundation, isolation cream, loose powder, air cushion, etc.), it can be associated with at least one of the cheek area, forehead area and jaw area of ​​the human face, and further, the skin features may include one or more of skin color features, texture features, pore features, stratum corneum features, and scar features.

[0076] Further references Figure 4 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a commodity recommendation device. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0077] like Figure 4 As shown, the commodity recommendation device 400 of the embodiment of the present disclosure may include: a determination module 401, an extraction module 402 and a push module 403. The determination module 401 is configured to determine the type of commodity to be recommended from the commodity recommendation request; the extraction module 402 is configured to extract the skin features corresponding to the type of commodity to be recommended from the face image of the target user; the push module 403 is configured to push the commodity information of the type of commodity to be recommended to the target user based on the skin features corresponding to the type of commodity to be recommended.

[0078] In the embodiment of the present disclosure, the specific processing of the determination module 401, the extraction module 402 and the push module 403 and the technical effects thereof can be referred to in Figure 2 The relevant descriptions of steps 201 - 203 in the corresponding embodiment are not repeated here.

[0079] In some optional implementations of the embodiments of the present disclosure, the skin features corresponding to the above-mentioned product types to be recommended include skin color features; and the above-mentioned push module 403 may further include a first determination submodule 4031, a second determination submodule 4032, and a push submodule 4033. The first determination submodule 4031 is configured to determine a first skin color feature vector based on the skin color features corresponding to the product types to be recommended; the second determination submodule 4032 is configured to determine similar users of the target user from the candidate users based on the skin color similarity between the first skin color feature vector and the second skin color feature vector of the candidate user; the push submodule 4033 is configured to push product information to the target user based on the first historical product usage information of the similar users; wherein the first historical product usage information corresponds to the product types to be recommended.

[0080] In the embodiment of the present disclosure, the specific processing of the first determination submodule 4031, the second determination submodule 4032 and the push submodule 4033 and the technical effects thereof can be referred to in Figure 3 The relevant descriptions of steps 303 - 305 in the corresponding embodiment are not repeated here.

[0081] In some optional implementations of the embodiments of the present disclosure, the push submodule 4033 is further configured to: determine a first product usage feature vector of a similar user based on the first historical product usage information; determine a second product usage feature vector of the target user based on the second historical product usage information of the target user; wherein the second historical product usage information corresponds to the type of product to be recommended; and determine the product information to be pushed to the target user from the first historical product usage information based on the similarity between the first product usage feature vector and the second product usage feature vector.

[0082] In some optional implementations of the embodiments of the present disclosure, the above-mentioned first determination submodule 4031 is further configured to: convert the first color space data of the skin color characteristics corresponding to the type of product to be recommended into second color space data; cluster the second color space data to obtain a first skin color feature vector.

[0083] In some optional implementations of the embodiments of the present disclosure, the above-mentioned extraction module 402 is further configured to determine the target facial area corresponding to the type of product to be recommended from the facial image of the target user based on the association relationship between the preset product type and the facial area of ​​the human face; and extract skin features corresponding to the type of product to be recommended based on the facial features of the detected target facial area.

[0084] In some optional implementations of the embodiments of the present disclosure, the above-mentioned types of goods to be recommended include at least one of the following: skin care products, cosmetics.

[0085] The disclosed embodiment exists as a device embodiment corresponding to the above method embodiment. The product recommendation device 400 provided by the disclosed embodiment needs to determine the type of product to be recommended from the user's product recommendation request, and extract the user's skin features that are compatible with the type of product to be recommended based on the user's face image, and then push the product information of the type of product to be recommended to the user based on the extracted skin features. In this way, a product recommendation method that is compatible with the user's product recommendation needs and takes into account the user's skin features can be automatically implemented, which is not limited to the user's personal aesthetics or preferences, but can meet the user's needs, thereby improving the accuracy of product recommendations.

[0086] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the commodity recommendation method described in any of the above embodiments can be implemented when the at least one processor executes.

[0087] According to an embodiment of the present disclosure, the present disclosure further provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to implement the product recommendation method described in any of the above embodiments when executed.

[0088] According to an embodiment of the present disclosure, the present disclosure further provides a computer program product, including a computer program, which can implement the commodity recommendation method described in any of the above embodiments when executed by a processor.

[0089] Figure 5 A schematic block diagram of an example electronic device 500 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0090] like Figure 5As shown, the device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0091] A number of components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0092] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above, such as a commodity recommendation method. For example, in some embodiments, the commodity recommendation method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the commodity recommendation method described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the commodity recommendation method in any other appropriate manner (e.g., by means of firmware).

[0093] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0094] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0095] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0096] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0097] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0098] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and virtual private servers (VPS) services.

[0099] According to the technical solution of the embodiment of the present disclosure, it is necessary to determine the type of product to be recommended from the user's product recommendation request, and extract the user's skin features that are compatible with the type of product to be recommended based on the user's face image, and then push the product information of the type of product to be recommended to the user based on the extracted skin features. In this way, a product recommendation method that is compatible with the user's product recommendation needs and takes into account the user's skin features can be automatically implemented, which is not limited to the user's personal aesthetics or preferences, but can meet the user's needs, thereby improving the accuracy of product recommendations.

[0100] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0101] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A product recommendation method, comprising: Determine the type of product to be recommended from the product recommendation request; Extracting skin features corresponding to the type of product to be recommended from the face image of the target user; Based on the skin features corresponding to the type of product to be recommended, product information of the type of product to be recommended is pushed to the target user.

2. The method according to claim 1, wherein: The skin features corresponding to the product type to be recommended include skin color features; as well as The pushing the commodity information of the commodity type to be recommended to the target user based on the skin feature corresponding to the commodity type to be recommended includes: Determining a first skin color feature vector based on the skin color feature corresponding to the type of product to be recommended; Determining a user similar to the target user from the candidate users based on the skin color similarity between the first skin color feature vector and the second skin color feature vector of the candidate user; Based on the first historical commodity usage information of the similar user, the commodity information is pushed to the target user; wherein the first historical commodity usage information corresponds to the type of the commodity to be recommended.

3. The method according to claim 2, wherein: The pushing the product information to the target user based on the first historical product usage information of the similar user includes: Determining a first product usage feature vector of the similar user based on the first historical product usage information; Determining a second commodity usage feature vector of the target user based on second historical commodity usage information of the target user; wherein the second historical commodity usage information corresponds to the type of commodity to be recommended; The product information pushed to the target user is determined from the first historical product usage information based on the similarity between the first product usage feature vector and the second product usage feature vector.

4. The method according to claim 2, wherein: The determining of the first skin color feature vector based on the skin color feature corresponding to the type of product to be recommended includes: Converting first color space data of skin color characteristics corresponding to the type of product to be recommended into second color space data; Clustering is performed on the second color space data to obtain the first skin color feature vector.

5. The method according to claim 1, wherein: The step of extracting skin features corresponding to the type of product to be recommended from the face image of the target user includes: Based on a preset association relationship between commodity types and facial regions of human faces, determining a target facial region corresponding to the commodity type to be recommended from the facial image of the target user; Based on the detected facial features of the target facial area, skin features corresponding to the type of product to be recommended are extracted.

6. The method according to any one of claims 1 to 5, wherein: The types of commodities to be recommended include at least one of the following: skin care products and cosmetics.

7. A commodity recommendation device, comprising: A determination module is configured to determine the type of product to be recommended from the product recommendation request; An extraction module is configured to extract skin features corresponding to the type of product to be recommended from the face image of the target user; The push module is configured to push product information of the type of product to be recommended to the target user based on the skin features corresponding to the type of product to be recommended.

8. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the commodity recommendation method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the product recommendation method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the steps of the product recommendation method according to any one of claims 1 to 6 are implemented.

Images