CN113033413A

CN113033413A - Glasses recommendation method and device, storage medium and terminal

Info

Publication number: CN113033413A
Application number: CN202110326501.9A
Authority: CN
Inventors: 唐俊锋
Original assignee: Xian Wingtech Information Technology Co Ltd
Current assignee: Xian Wingtech Information Technology Co Ltd
Priority date: 2021-03-26
Filing date: 2021-03-26
Publication date: 2021-06-25

Abstract

The invention discloses a glasses recommendation method, a device, a storage medium and a terminal, wherein the method comprises the following steps: acquiring a target face picture of a user to be recommended; detecting the target face picture to obtain a first target face feature and a second target face feature; generating an initial eyewear style that matches the first target facial feature; simulating the wearing effect of the initial glasses style on the target face picture by using an AR technology; acquiring a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to a second target facial feature; the best matching target eyewear style and its purchase link are generated. The invention can provide accurate matching picture frame service for users by utilizing the face detection technology and the AR technology, myopia people can select a proper picture frame through the scheme, and not only can provide a product with higher cost performance for the myopia people through comparison with an offline glasses shop, but also greatly reduces the risk of improper goods return when the glasses are matched on line.

Description

Glasses recommendation method and device, storage medium and terminal

[ technical field ] A method for producing a semiconductor device

The invention relates to the field of computers, in particular to a glasses recommendation method, a glasses recommendation device, a storage medium and a terminal.

[ background of the invention ]

In daily life, glasses are necessities of myopia crowds, people select a glasses frame and perform optometry to match glasses through an offline glasses shop, and more people select to purchase glasses on the internet along with the rise of online shopping.

The biggest advantage of online shopping is that the user can not select the commodity of own mood apparatus, the glasses are not exceptional, and the general online shop can provide information such as mirror surface width, bridge width (nose bridge width), frame height, whole frame length (frame face outer width), mirror leg length of frame, combines the degree of a consumer's own eyes, interpupillary distance, etc., just can carry out the frame shopping in the online shop. However, the simple matching purchasing method is difficult for consumers to purchase satisfactory glasses on the internet, and consumers who have glasses fitting experience know that the users need to debug the glasses on the face after the glasses are fitted due to different pupil distances, different bridge heights and the like of the users, otherwise, the users are very uncomfortable to wear and even influence the eyesight of the glasses fittings.

[ summary of the invention ]

The invention provides a glasses recommendation method, a glasses recommendation device, a storage medium and a terminal, and solves the technical problem that the throughput rate test result in the prior art is inaccurate.

The technical scheme for solving the technical problems is as follows: an eyeglass recommendation method comprising the steps of:

acquiring a target face picture of a user to be recommended;

detecting the target facial picture, and acquiring a first target facial feature for representing wearing aesthetic degree and a second target facial feature for representing wearing comfort;

generating at least one initial eyewear style matching the first target facial feature using the trained deep neural network;

simulating the wearing effect of each initial glasses style on the target face picture through an AR technology;

when the user to be recommended selects an optimal wearing effect from all wearing effects, acquiring a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to the second target facial feature;

and generating a best matched target glasses style and a purchasing link thereof according to the first target glasses parameter and the second target glasses parameter so as to be selected and purchased by the user to be recommended.

In a preferred embodiment, the selecting the optimal wearing effect from all wearing effects comprises:

continuously acquiring a voice improvement instruction and/or a character improvement instruction of the user to be recommended aiming at any wearing effect, performing semantic analysis on the voice improvement instruction and/or the character improvement instruction, and adjusting the current glasses parameters corresponding to the initial glasses style according to a semantic analysis result until the user to be recommended confirms the optimal wearing effect.

In a preferred embodiment, the first target facial feature includes at least one of a hairstyle, an eyebrow shape, a facial proportion, an eyebrow space, and an orbital width height of the user to be recommended;

the second target facial features comprise at least one of the bridge height, the bridge width, the side face width and the interpupillary distance of the user to be recommended;

the first target eyeglass parameter comprises at least one of a mirror surface width, a frame shape, a frame width, a frame height, a frame thickness, a frame outer width, a temple shape and a frame material;

the second target eyewear parameters include a bridge width and/or a temple length.

In a preferred embodiment, the generating at least one initial eyeglass style matching the first target facial feature using the trained deep neural network comprises:

constructing a glasses style recommendation model based on a deep neural network, wherein the input of the glasses style recommendation model at least comprises a first facial feature, and the output of the glasses style recommendation model is at least one initial glasses style;

acquiring a sample picture set, and screening out training pictures wearing glasses from the sample picture set;

training the glasses style recommendation model by adopting the training picture, and optimizing model parameters of the glasses style recommendation model;

and generating at least one initial glasses style matched with the first target facial features of the user to be recommended by adopting the trained glasses style recommendation model.

In a preferred embodiment, the input of the glasses style recommendation model further comprises demographic characteristics, the demographic characteristics comprising at least one of user age, gender, occupation and glasses usage occasion.

A second aspect of an embodiment of the present invention provides an eyeglass recommendation method, including the steps of:

acquiring target face pictures of a plurality of users to be recommended;

detecting each target face picture to obtain a first target face feature and a second target face feature corresponding to each user to be recommended;

generating at least one initial eyeglass pattern matched with all the first target facial features by adopting the trained deep neural network;

simultaneously displaying target face pictures of all users to be recommended, and simulating the wearing effect of the same initial glasses style on the corresponding target face pictures by using an AR (augmented reality) technology;

when all users to be recommended select the same optimal wearing effect, acquiring a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to a second target facial feature of each user to be recommended;

and generating a best matched target glasses style and a purchasing link thereof for each user to be recommended according to the first target glasses parameter and the second target glasses parameter so as to be selected and purchased by the user to be recommended.

A third aspect of the embodiments of the present invention provides an eyeglass recommendation apparatus, including a first obtaining module, a first detecting module, a first matching module, a first display module, a second obtaining module, and a first recommending module,

the first acquisition module is used for acquiring a target face picture of a user to be recommended;

the first detection module is used for detecting the target facial picture and acquiring a first target facial feature for representing wearing aesthetic degree and a second target facial feature for representing wearing comfort;

the first matching module is used for generating at least one initial glasses style matched with the first target facial feature by adopting the trained deep neural network;

the first display module is used for simulating the wearing effect of each initial glasses style on the target facial picture through an AR technology;

the second obtaining module is used for obtaining a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to the second target facial feature after the user to be recommended selects the optimal wearing effect from all the wearing effects;

the first recommending module is used for generating a best matched target glasses style and a purchasing link thereof according to the first target glasses parameter and the second target glasses parameter so as to be selected and purchased by the user to be recommended.

A fourth aspect of the embodiments of the present invention provides an eyeglass recommendation apparatus, including a third obtaining module, a second detecting module, a second matching module, a second displaying module, a fourth obtaining module, and a second recommending module,

the third acquisition module is used for acquiring target face pictures of a plurality of users to be recommended;

the second detection module is used for detecting each target face picture to obtain a first target face feature and a second target face feature corresponding to each user to be recommended;

the second matching module is used for generating at least one initial glasses pattern matched with all the first target facial features by adopting the trained deep neural network;

the second display module is used for simultaneously displaying the target face pictures of all the users to be recommended and simulating the wearing effect of the same initial glasses style on the corresponding target face pictures through an AR technology;

the fourth obtaining module is used for obtaining a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to a second target facial feature of each user to be recommended after all users to be recommended select the same optimal wearing effect;

the second recommending module is used for generating a best matched target glasses style and a purchasing link thereof for each user to be recommended according to the first target glasses parameter and the second target glasses parameter so as to be selected and purchased by the user to be recommended.

A fifth aspect of embodiments of the present invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the glasses recommendation method described in any one of the above.

A sixth aspect of the embodiments of the present invention provides a terminal, including the computer-readable storage medium and a processor, where the processor implements the above-mentioned steps of the glasses recommendation method when executing the computer program on the computer-readable storage medium.

The invention provides a glasses recommendation method, a device, a storage medium and a terminal, which utilize a face detection technology and an AR technology, can facilitate a user to select the style and style of glasses according to a first target facial feature of the user, and then adjust the detail size of the glasses of the selected style or style according to a second target facial feature of the user, thereby obtaining the most appropriate target glasses style and purchasing link thereof, ensuring the attractiveness and comfort of glasses wearing, and providing accurate glasses frame matching service for the user. Compared with an offline glasses shop, the online glasses matching system can provide a product with higher cost performance for myopia crowds, meanwhile, the risk of improper goods return and exchange of online glasses matching is greatly reduced, the collected facial information data of the user can be used for matching a picture frame again in the future, and can be matched with optometry data, so that online integrated glasses matching is achieved, and the online glasses matching system has extremely high popularization prospect.

In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.

[ description of the drawings ]

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.

FIG. 1 is a diagram of an exemplary implementation of a method for glasses recommendation;

FIG. 2 is a schematic flow chart diagram of a method for glasses recommendation in one embodiment;

FIG. 3 is a flowchart illustrating a method for recommending glasses according to another embodiment;

FIG. 4 is a schematic diagram of the structure of the glasses recommendation device in one embodiment;

FIG. 5 is a schematic diagram of the structure of an eyeglass recommender in another embodiment;

FIG. 6 is a diagram illustrating an internal structure of a computer device in one embodiment.

[ detailed description ] embodiments

In order to make the objects, technical solutions and advantageous effects of the present invention more clearly apparent, the present invention is further described in detail below with reference to the accompanying drawings and the detailed description. It should be understood that the detailed description and specific examples, while indicating the preferred embodiment of the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention.

The glasses recommendation method provided by the application can be applied to the application environment shown in fig. 1. The glasses recommendation method is applied to the terminal 102, wherein the terminal 102 may be, but not limited to, a smart phone, a computer, a notebook computer, a palm computer, a Personal Digital Assistant (PDA), a Portable Media Player (PMP), a wearable device, a smart band, and other mobile terminals, and fixed terminals such as a digital TV, a desktop computer, and the like.

In one embodiment, as shown in fig. 2, an eyeglass recommendation method is provided, which is described by taking the example that the method is applied to the smartphone in fig. 1, and includes the following steps:

and S11, the smart phone acquires a target face picture of the user to be recommended. In a preferred embodiment, the acquiring, by the smartphone, the target face picture of the user to be recommended includes:

and S110, the smart phone receives a user instruction and starts the camera shooting unit according to the user instruction.

And S111, the smart phone collects real-time facial pictures of the user to be recommended through the camera unit, wherein the real-time facial pictures comprise at least one front picture and at least one side picture. In order to improve the accuracy of the facial features of the user, the corresponding head portrait frame can be displayed in the shooting process, so that the user is instructed to put the face into the head portrait frame for shooting, and the user is instructed to adjust the position of the face in the head portrait frame during shooting until the whole face or part of the face of the user falls into the head portrait frame. In other embodiments, the facial photos saved in the user's smartphone may also be obtained directly.

And S112, selecting an easily-identified target front picture and a target side picture from the at least one front picture and the at least one side picture according to parameters such as light, exposure and/or definition, so as to obtain a target face picture of the user to be recommended, wherein the target face picture may include one target front picture and one target side picture.

And then S12 is executed, the smart phone detects the target facial picture, and a first target facial feature for representing wearing aesthetic degree and a second target facial feature for representing wearing comfort degree are obtained. Generally, the glasses brands of online stores or online shopping malls are very many, different brands usually have glasses frames with basically the same or similar styles, and the wearing attractiveness of the glasses frames with similar styles or styles on the face is not greatly different, but the sizes of the specific glasses frames are slightly different, so that the wearing comfort or the price of the glasses are influenced, and the purchasing tendency of the glasses is influenced. Therefore, the present embodiment divides the facial features of the user into a first target facial feature and a second target facial feature, the first target facial feature is mainly related to the style and style of the frame to determine the aesthetic degree of the glasses wearing, and the second target facial feature is mainly related to the wearing size of the frame to determine the comfort degree of the glasses wearing. Specifically, the first target facial feature specifically includes at least one of a hairstyle, an eyebrow shape, a facial proportion, an eyebrow space, and an orbital width height. The second target facial feature includes at least one of a nose bridge height, a nose bridge width, a side face width, and a pupillary distance.

In a preferred embodiment, the detection of the target face picture by the smartphone comprises the following steps:

s121, detecting a target face picture by the smart phone by adopting a preset face recognition algorithm, and acquiring coordinates of at least one preset face feature point;

and S122, the smart phone calculates corresponding first target facial features and second target facial features according to the coordinates of at least one preset facial feature point.

The preset face recognition algorithm mainly adopts a deep convolutional neural network to extract features, namely the trained deep convolutional neural network can directly recognize numerical values such as hairstyle, eyebrow shape, eyebrow spacing, eye socket width and height, nose bridge width, nose bridge height, side face width, pupil distance, face proportion and the like of a target face picture. A deep Convolutional Neural Network (CNN) is a kind of feedforward neural network, and generally, the basic structure of the CNN includes two layers, one of which is a feature extraction layer, and the input of each neuron is connected to a local acceptance domain of the previous layer and extracts the feature of the local. Once the local feature is extracted, the position relation between the local feature and other features is determined; the other is a feature mapping layer, each calculation layer of the network is composed of a plurality of feature mappings, each feature mapping is a plane, and the weights of all neurons on the plane are equal. The feature mapping structure adopts a sigmoid function with small influence function kernel as an activation function of the convolution network, so that the feature mapping has displacement invariance. In addition, since the neurons on one mapping surface share the weight, the number of free parameters of the network is reduced. Each convolutional layer in the deep convolutional neural network is followed by a computation layer for local averaging and quadratic extraction, and the characteristic quadratic feature extraction structure reduces the feature resolution. The advantage of training by using the deep convolutional neural network is that the deep convolutional neural network has a good effect in the field of image processing, and the training effect of the machine learning model can be improved. Based on the fact that the deep convolutional neural network technology is mature in the field of face recognition, the specific training method and the recognition process of the deep convolutional neural network are not explained in detail.

In another preferred embodiment, the target face picture may further include multiple target front pictures and multiple target side pictures, and therefore, a deep convolutional neural network technique may be used to perform feature extraction on all target face pictures, so as to obtain multiple facial feature data, so as to screen out erroneous data, such as data that is not within a preset value range or that obviously does not conform to the symmetry principle, and improve the accuracy of identification.

Then S13 is executed, and the smartphone generates at least one initial eyewear style matching the first target facial features using the trained deep neural network. The method specifically comprises the following steps:

s131, the smart phone constructs a glasses style recommendation model based on the deep neural network, the input of the glasses style recommendation model at least comprises first facial features, and the output of the glasses style recommendation model is at least one initial glasses style.

S132, the smart phone obtains a sample picture set, and the training pictures with the glasses are screened out from the sample picture set.

S133, the smart phone extracts the first training facial features and the training glasses style of the training pictures at first, then trains the glasses style recommendation model by adopting the first training facial features and the training glasses style of the training pictures, and optimizes model parameters of the glasses style recommendation model.

S134, the smart phone generates at least one initial glasses style matched with the first target facial features of the user to be recommended by the trained glasses style recommendation model.

In a preferred embodiment, the input to the eyewear style recommendation model further comprises demographic characteristics including at least one of user age, gender, expected purchase price, occupation, and eyewear usage. At this time, a basis function X may be set in the deep neural network, where Xa is a feature value of the feature a, a may be an influence weight of the feature a on the eyeglass style, for example, the feature a may be an eyebrow shape, similarly, the feature B may be a face shape, the feature C may be a hair style, and the feature D may be an age, and the like. The number of the features may be multiple, which is not specifically limited in this embodiment, and when a corresponding feature is not obtained in a picture, a feature value of the feature may be input to be 0. Meanwhile, the preferred embodiment may set different influence weights for each feature according to the importance of the feature on selecting the glasses style, for example, set a higher influence weight for the first target facial feature and set a reduced influence weight for the crowd feature, thereby generating the most appropriate initial glasses style.

Then S14 is executed, and the smart phone simulates wearing effect of each initial glasses style on the target facial picture through the AR technology. The method specifically comprises the following steps:

s141, the smart phone generates a corresponding glasses three-dimensional model according to the initial glasses style;

and S142, the smart phone synthesizes the three-dimensional model of the glasses and the target face picture of the user to be recommended by adopting an AR technology to generate a wearing effect picture. The preferred embodiment can generate the front wearing effect picture and the side wearing effect picture, and is convenient for a user to more visually see the wearing effect.

And then executing S15, and after the user to be recommended selects the optimal wearing effect from all the wearing effects, the smart phone obtains a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to the second target facial feature. The first target eyeglass parameters include at least one of a lens surface width, a frame shape, a frame width, a frame height, a frame thickness, an out-of-frame width, a temple shape, and a frame material, and are mainly related to a style and style of the eyeglasses and determined by a first target facial feature of the user. And the second target eyewear parameters include bridge width and/or temple length, primarily related to eyewear comfort, determined by the user's second target facial features.

In a preferred embodiment, the selecting the optimal wearing effect from all wearing effects comprises: the smart phone continuously obtains a voice improvement instruction and/or a text improvement instruction of a user to be recommended aiming at any wearing effect, for example, the user can say 'more beautiful,', 'the picture frame is changed into silver', 'the picture frame is changed into a circle' and the like, semantic analysis is carried out on the voice improvement instruction and/or the text improvement instruction, current glasses parameters corresponding to the initial glasses style are adjusted according to semantic analysis results, and the user to be recommended confirms the optimal wearing effect.

And then executing S16, and generating the best matched target glasses style and the purchase link thereof according to the first target glasses parameter and the second target glasses parameter for the selection and purchase of the user to be recommended.

The embodiment provides a glasses recommendation method, and by using a face detection technology and an AR technology, a user can select the style and style of glasses according to a first target facial feature of the user, and then adjust the detail size of the glasses in the selected style or style according to a second target facial feature of the user, so that the most suitable target glasses style and purchasing link of the glasses are obtained, the attractiveness and comfort of glasses wearing are guaranteed, and accurate glasses frame matching service is provided for the user. Compared with an offline glasses shop, the online glasses matching system can provide a product with higher cost performance for myopia crowds, meanwhile, the risk of improper goods return and exchange of online glasses matching is greatly reduced, the collected facial information data of the user can be used for matching a picture frame again in the future, and can be matched with optometry data, so that online integrated glasses matching is achieved, and the online glasses matching system has extremely high popularization prospect.

Another embodiment of the present invention provides another glasses recommendation method, as shown in fig. 3, including the following steps:

s21, obtaining target face pictures of a plurality of users to be recommended;

s22, detecting each target face picture, and acquiring a first target face feature and a second target face feature corresponding to each user to be recommended;

s23, generating at least one initial glasses pattern matched with all the first target facial features by adopting the trained deep neural network;

s24, simultaneously displaying target face pictures of all users to be recommended, and simulating the wearing effect of the same initial glasses style on the corresponding target face pictures through an AR technology;

s25, when all users to be recommended select the same optimal wearing effect, acquiring a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to a second target facial feature of each user to be recommended;

and S26, generating the best matched target glasses style and the purchasing link thereof for each user to be recommended according to the first target glasses parameter and the second target glasses parameter, so that the user to be recommended can select and purchase.

The glasses recommendation method mainly aims at multiple users who need to keep uniform dressing or want to buy glasses of the same style, firstly recommends an initial glasses style which can be adapted to each user according to the first target facial features of the users, namely recommends an initial glasses style which simultaneously ensures the wearing aesthetic degree of each user, and compared with a scheme of recommending glasses for one person, the matching degree can be set to be relatively low at the moment. And then obtaining the optimal wearing effect selected by each user, wherein the optimal wearing effect can be formed by fine tuning the same initial glasses style by different users, and meanwhile, the second target facial features of the users are combined to generate the best matched target glasses style and the purchasing link thereof for each user so as to be selected and purchased by the users, thereby meeting the purchasing requirements of the users.

In another embodiment, the method of the invention can be combined with eye optometry, namely, after the user is subjected to optometry by a smart phone, an optometry result is combined with a recommended result of the glasses frame, and optometry and frame selection are achieved in one step.

It should be noted that, for technical details that are not described in detail in the above embodiments of a method for recommending glasses for multiple users, reference may be made to the method for recommending glasses for one user provided in the embodiments of the present invention, and corresponding beneficial effects are achieved.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

Fig. 4 is a schematic structural diagram of an eyeglass recommendation apparatus according to another embodiment of the present invention, as shown in fig. 4, including a first obtaining module 100, a first detecting module 101, a first matching module 102, a first displaying module 103, a second obtaining module 104, and a first recommending module 105,

the first obtaining module 100 is configured to obtain a target face picture of a user to be recommended;

the first detection module 101 is configured to detect a target facial picture, and acquire a first target facial feature for representing wearing beauty and a second target facial feature for representing wearing comfort;

the first matching module 102 is configured to generate at least one initial eyeglass style matching the first target facial feature using the trained deep neural network;

the first display module 103 is used for simulating the wearing effect of each initial glasses style on the target facial picture through an AR technology;

the second obtaining module 104 is configured to obtain a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to a second target facial feature after the user to be recommended selects the optimal wearing effect from all wearing effects;

the first recommendation module 105 is configured to generate a best matching target glasses style and a purchase link thereof according to the first target glasses parameter and the second target glasses parameter, so that the user to be recommended selects and purchases the target glasses style.

In a preferred embodiment, the second obtaining module 104 is specifically configured to continuously obtain a voice improvement instruction and/or a text improvement instruction of the user to be recommended for any wearing effect, perform semantic analysis on the voice improvement instruction and/or the text improvement instruction, and adjust the current glasses parameters corresponding to the initial glasses style according to a semantic analysis result until the user to be recommended confirms the optimal wearing effect.

In a preferred embodiment, the first target facial features include at least one of a hairstyle, an eyebrow shape, a facial proportion, an eyebrow space, and an orbital width height of the user to be recommended;

the second target eyewear parameters include bridge width and/or temple length.

In a preferred embodiment, the first matching module 102 specifically includes:

the glasses model recommendation system comprises a construction unit, a recommendation unit and a recommendation unit, wherein the construction unit is used for constructing a glasses model recommendation model based on a deep neural network, the input of the glasses model recommendation model at least comprises first facial features, and the output of the glasses model recommendation model is at least one initial glasses model;

the sample acquisition unit is used for acquiring a sample picture set and screening out a training picture wearing glasses from the sample picture set;

the training unit is used for training the glasses style recommendation model by adopting the training pictures and optimizing model parameters of the glasses style recommendation model;

and the matching unit is used for generating at least one initial glasses style matched with the first target facial features of the user to be recommended by adopting the trained glasses style recommendation model.

The embodiment provides a glasses recommending device, and by using a face detection technology and an AR technology, a user can select the style and style of glasses according to a first target facial feature of the user, and then adjust the detail size of the glasses in the selected style or style according to a second target facial feature of the user, so that the most suitable target glasses style and purchasing link of the glasses are obtained, the attractiveness and comfort of glasses wearing are guaranteed, and accurate glasses frame matching service is provided for the user. Compared with an offline glasses shop, the online glasses matching system can provide a product with higher cost performance for myopia crowds, meanwhile, the risk of improper goods return and exchange of online glasses matching is greatly reduced, the collected facial information data of the user can be used for matching a picture frame again in the future, and can be matched with optometry data, so that online integrated glasses matching is achieved, and the online glasses matching system has extremely high popularization prospect.

The glasses recommendation device can execute the glasses recommendation method provided by the embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details that are not described in detail in the embodiments of the glasses recommendation device, reference may be made to the glasses recommendation method provided in the embodiments of the present invention.

Another embodiment of the present invention provides another glasses recommending apparatus, as shown in fig. 5, including a third obtaining module 200, a second detecting module 201, a second matching module 202, a second displaying module 203, a fourth obtaining module 204, and a second recommending module 205,

the third obtaining module 200 is configured to obtain target facial pictures of a plurality of users to be recommended;

the second detection module 201 is configured to detect each target facial picture, and obtain a first target facial feature and a second target facial feature corresponding to each user to be recommended;

the second matching module 202 is configured to generate at least one initial eyeglass style that matches all of the first target facial features using the trained deep neural network;

the second display module 203 is configured to display the target facial pictures of all the users to be recommended at the same time, and simulate the wearing effect of the same initial glasses style on the corresponding target facial picture through an AR technology;

the fourth obtaining module 204 is configured to obtain, after all users to be recommended select the same optimal wearing effect, a first target eyeglass parameter corresponding to the optimal wearing effect and a second target eyeglass parameter corresponding to a second target facial feature of each user to be recommended;

the second recommendation module 205 is configured to generate a best-matched target glasses style and a purchase link thereof for each user to be recommended according to the first target glasses parameter and the second target glasses parameter, so that the user to be recommended can select and purchase the best-matched target glasses style and the purchase link thereof.

The glasses recommendation device is mainly used for multiple users who need to keep uniform dressing or want to buy glasses of the same style, and firstly recommends an initial glasses style which can be adapted to each user according to the first target facial features of the users, namely recommends an initial glasses style which simultaneously ensures the wearing attractiveness of each user. And then obtaining the optimal wearing effect selected by each user, wherein the optimal wearing effect can be formed by fine tuning the same initial glasses style by different users, and meanwhile, the second target facial features of the users are combined to generate the best matched target glasses style and the purchasing link thereof for each user so as to be selected and purchased by the users, thereby meeting the purchasing requirements of the users.

It should be noted that, for technical details that are not described in detail in the above embodiments of an apparatus for recommending glasses for multiple users, reference may be made to the apparatus for recommending glasses for a user according to the embodiments of the present invention, and corresponding advantageous effects are achieved.

In one embodiment, the invention also provides a computer-readable storage medium having a computer program stored thereon, the computer program, when executed by a processor, implementing the steps of:

s11, acquiring a target face picture of the user to be recommended;

s12, detecting the target face picture, and acquiring a first target face feature for representing the wearing aesthetic degree and a second target face feature for representing the wearing comfort degree;

s13, generating at least one initial glasses pattern matched with the first target facial feature by adopting the trained deep neural network;

s14, simulating the wearing effect of each initial glasses style on the target face picture through AR technology;

s15, when the user to be recommended selects the optimal wearing effect from all the wearing effects, acquiring a first target glasses parameter corresponding to the optimal wearing effect and a second target glasses parameter corresponding to a second target facial feature;

and S16, generating the best matched target glasses style and the purchasing link thereof according to the first target glasses parameter and the second target glasses parameter, so as to be selected and purchased by the user to be recommended.

In one embodiment, the computer program when executed by the processor further performs the steps of:

continuously acquiring a voice improvement instruction and/or a character improvement instruction of a user to be recommended aiming at any wearing effect, performing semantic analysis on the voice improvement instruction and/or the character improvement instruction, and adjusting the current glasses parameters corresponding to the initial glasses style according to the semantic analysis result until the user to be recommended confirms the optimal wearing effect.

In one embodiment, the computer program, when executed by the processor, includes at least one of a hairstyle, an eyebrow shape, a facial proportion, an eyebrow spacing, and an orbital width height of the user to be recommended;

the second target eyewear parameters include bridge width and/or temple length.

s131, constructing a glasses style recommendation model based on a deep neural network, wherein the input of the glasses style recommendation model at least comprises first facial features, and the output of the glasses style recommendation model is at least one initial glasses style;

s132, acquiring a sample picture set, and screening out training pictures wearing glasses from the sample picture set;

s133, training the glasses style recommendation model by adopting the training pictures, and optimizing model parameters of the glasses style recommendation model;

and S134, generating at least one initial glasses style matched with the first target facial features of the user to be recommended by adopting the trained glasses style recommendation model.

In one embodiment, the input of the eyewear style recommendation model when the computer program is executed by the processor further comprises demographic characteristics, the demographic characteristics comprising at least one of user age, gender, occupation, and eyewear usage.

The above embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the glasses recommendation method is executed, so that a user can select a style and a style of glasses according to a first target facial feature of the user and then adjust a detail size of the selected style or style of glasses according to a second target facial feature of the user by using a face detection technology and an AR technology, so as to obtain a most suitable target glasses style and a purchasing link thereof, thereby ensuring the beauty and comfort of glasses wearing, and providing a precise glasses frame selection service for the user. Compared with an offline glasses shop, the online glasses matching system can provide a product with higher cost performance for myopia crowds, meanwhile, the risk of improper goods return and exchange of online glasses matching is greatly reduced, the collected facial information data of the user can be used for matching a picture frame again in the future, and can be matched with optometry data, so that online integrated glasses matching is achieved, and the online glasses matching system has extremely high popularization prospect.

s21, obtaining target face pictures of a plurality of users to be recommended;

The above embodiment proposes a computer-readable storage medium on which a computer program is stored, the computer program being executed by a processor to perform a glasses recommendation method, when there are a plurality of users who need to keep uniform dressing or want to purchase glasses of the same style, first recommending an initial glasses style that can be fitted to each person according to their first target facial features, that is, recommending an initial glasses style that ensures the wearing beauty of each person at the same time, when the degree of matching can be set relatively low compared to a scheme in which glasses recommendation is made for one person. And then obtaining the optimal wearing effect selected by each user, wherein the optimal wearing effect can be formed by fine tuning the same initial glasses style by different users, and meanwhile, the second target facial features of the users are combined to generate the best matched target glasses style and the purchasing link thereof for each user so as to be selected and purchased by the users, thereby meeting the purchasing requirements of the users.

Fig. 6 is an internal structure diagram of a computer device in an embodiment, and the computer device may be a notebook computer, or may be another mobile terminal or a fixed terminal. As shown in fig. 6, the apparatus comprises a memory 81 and a processor 80, the memory 81 stores a computer program 82, and the processor 80 implements the following steps when executing the computer program 82:

s11, acquiring a target face picture of the user to be recommended;

In one embodiment, the processor 80, when executing the computer program 82, further performs the steps of:

In one embodiment, the input to the eyewear style recommendation model when the processor 80 executes the computer program 82 further includes demographic characteristics including at least one of user age, gender, occupation, and eyewear usage.

In one embodiment, when the processor 80 executes the computer program 82, the first target facial features include at least one of a hairstyle, an eyebrow shape, a facial proportion, an eyebrow spacing, and an orbital width height of the user to be recommended;

the second target eyewear parameters include bridge width and/or temple length.

The embodiment utilizes the face detection technology and the AR technology, so that a user can select the style and style of the glasses conveniently according to the first target facial feature of the user, and then adjust the detailed size of the glasses in the selected style or style according to the second target facial feature of the user, so that the most appropriate target glasses style and purchasing link of the glasses are obtained, the attractiveness and comfort of glasses wearing are ensured, and accurate glasses frame selecting service is provided for the user. Compared with an offline glasses shop, the online glasses matching system can provide a product with higher cost performance for myopia crowds, meanwhile, the risk of improper goods return and exchange of online glasses matching is greatly reduced, the collected facial information data of the user can be used for matching a picture frame again in the future, and can be matched with optometry data, so that online integrated glasses matching is achieved, and the online glasses matching system has extremely high popularization prospect.

In another preferred embodiment, the processor 80, when executing the computer program 82, further performs the steps of:

s21, obtaining target face pictures of a plurality of users to be recommended;

When there are a plurality of users who need to keep wearing the same or want to purchase the same style of glasses in a unified manner, the above-described embodiment first recommends an initial glasses style that can be adapted to each person according to their first target facial features, that is, recommends an initial glasses style that ensures the wearing beauty of each person at the same time, and at this time, the degree of matching can be set relatively low compared to a scheme in which glasses recommendation is made for one person. And then obtaining the optimal wearing effect selected by each user, wherein the optimal wearing effect can be formed by fine tuning the same initial glasses style by different users, and meanwhile, the second target facial features of the users are combined to generate the best matched target glasses style and the purchasing link thereof for each user so as to be selected and purchased by the users, thereby meeting the purchasing requirements of the users.

It will be understood by those skilled in the art that fig. 6 is only one example of the terminal of the present invention, and is not limited to the terminal, and may include more or less components than those shown, or combine some components, or different components, for example, the terminal may further include a power management module, an arithmetic processing module, an input/output device, a network access device, a bus, etc.

The Processor 80 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The storage 81 may be an internal storage unit of the terminal, such as a hard disk or a memory. The memory 81 may also be an external storage device of the terminal, such as a plug-in hard disk provided on the compass calibration terminal, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like. Further, the memory 81 may also include both an internal storage unit of the compass calibration terminal and an external storage device. The memory 81 is used to store computer programs and other programs and data needed for compass calibration of the terminal. The memory 81 may also be used to temporarily store data that has been output or is to be output.

It will be apparent to those skilled in the art that, for convenience and simplicity of description, the foregoing functional units and modules are merely illustrated in terms of division, and in practical applications, the foregoing functional allocation may be performed by different functional units and modules as needed, that is, the internal structure of the terminal is divided into different functional units or modules to perform all or part of the above described functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the above-mentioned apparatus refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and method steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the embodiments provided in the present invention, it should be understood that the disclosed terminal/terminal device and method can be implemented in other ways. For example, the above-described terminal/terminal device embodiments are merely illustrative, and for example, a module or a unit may be divided into only one logic function, and may be implemented in other ways, for example, a plurality of units or components may be combined or integrated into another device, or some features may be omitted or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, terminals or units, and may be in an electrical, mechanical or other form.

Units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The invention is not limited solely to that described in the specification and embodiments, and additional advantages and modifications will readily occur to those skilled in the art, so that the invention is not limited to the specific details, representative apparatus, and illustrative examples shown and described herein, without departing from the spirit and scope of the general concept as defined by the appended claims and their equivalents.

Claims

1. An eyeglass recommendation method, comprising the steps of:

acquiring a target face picture of a user to be recommended;

2. The eyeglass recommendation method according to claim 1, wherein selecting the optimal wearing effect from all wearing effects comprises:

3. The eyewear recommendation method of claim 1, wherein the first target facial feature comprises at least one of a hairstyle, an eyebrow shape, a face proportion, an eyebrow spacing, and an orbital width height of a user to be recommended;

4. The eyewear recommendation method of any of claims 1-3, wherein the generating at least one initial eyewear style matching the first target facial feature using the trained deep neural network comprises:

5. The eyewear recommendation method of claim 4, wherein the input of the eyewear style recommendation model further comprises demographic characteristics, the demographic characteristics comprising at least one of user age, gender, occupation, and eyewear usage.

6. An eyeglass recommendation method, comprising the steps of:

acquiring target face pictures of a plurality of users to be recommended;

7. A spectacle recommendation device is characterized by comprising a first acquisition module, a first detection module, a first matching module, a first display module, a second acquisition module and a first recommendation module,

8. A glasses recommending device is characterized by comprising a third obtaining module, a second detecting module, a second matching module, a second display module, a fourth obtaining module and a second recommending module,

9. A computer-readable storage medium, in which a computer program is stored, which, when executed by a processor, implements the glasses recommendation method according to any one of claims 1 to 7.

10. A terminal, characterized in comprising the computer-readable storage medium of claim 9 and a processor which, when executing a computer program on the computer-readable storage medium, carries out the steps of the glasses recommendation method of any of claims 1-7.