Image construction method, image construction device, electronic device and storage medium

Abstract

An embodiment of the present application provides an image construction method, an image construction device, an electronic device, and a storage medium, belonging to the field of artificial intelligence technology. The method includes: obtaining target growth data, where the target growth data is the data of the learning growth of a target user; performing data filtering processing on the target growth data to obtain target data; performing prediction processing on the target data through a preset label prediction model to obtain target image labels of the target user; calculating weights for preset image attributes according to the target image labels to obtain weight values corresponding to each preset image attribute; calculating a growth score for the target user according to the weight values to obtain a comprehensive score value; and performing image construction processing on the preset image attributes according to the weight values and the comprehensive score value to obtain a target user image. The embodiment of the present application can improve the accuracy of image construction.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and particularly to an image construction method, an image construction device, an electronic device, and a storage medium. Background Art

[0002] Currently, most image construction methods rely on the basic data of the target user to construct the user image, but the basic data often cannot cover multiple business scenarios, which often causes the problem of incomplete images, thus affecting the accuracy of image construction. Therefore, how to improve the accuracy of image construction has become a technical problem to be solved urgently. Summary of the Invention

[0003] The main purpose of the embodiments of this application is to propose an image construction method, an image construction device, an electronic device, and a storage medium, aiming to improve the accuracy of image construction.

[0004] To achieve the above object, in the first aspect of the embodiments of this application, an image construction method is proposed. The method includes:

[0005] Obtain target growth data, where the target growth data is the learning growth data of the target user;

[0006] Perform data filtering processing on the target growth data to obtain target data;

[0007] Perform prediction processing on the target data through a preset label prediction model to obtain the target image labels of the target user;

[0008] Calculate the weights of the preset image attributes according to the target image labels to obtain the weight values corresponding to each preset image attribute;

[0009] Calculate the growth score of the target user according to the weight values to obtain a comprehensive score value;

[0010] Perform image construction processing on the preset image attributes according to the weight values and the comprehensive score value to obtain the target user image.

[0011] In some embodiments, the label prediction model includes an embedding layer, a convolutional layer, a clustering layer, and a prediction layer. The step of performing prediction processing on the target data through the preset label prediction model to obtain the target image labels of the target user includes:

[0012] Perform word embedding processing on the target data through the embedding layer to obtain a target embedding vector;

[0013] Perform entity feature extraction on the target embedding vector through the convolutional layer to obtain target entity features;

[0014] Integrate and process the target entity features through the clustering layer to obtain central entity features;

[0015] Calculate the label probability of the central entity features through a preset function of the prediction layer to obtain a label probability value;

[0016] Determine the target portrait label according to the label probability value.

[0017] In some embodiments, the step of calculating the weight value corresponding to each preset portrait attribute according to the target portrait label includes:

[0018] Obtain a preset relationship mapping table, which stores preset portrait labels and preset portrait attributes, and there is a mapping relationship between the preset portrait labels and the preset portrait attributes;

[0019] Extract the target portrait label from the preset portrait labels;

[0020] Extract the mapping data between the target portrait label and the preset portrait attributes from the relationship mapping table;

[0021] Calculate the weight of the preset portrait attributes according to the mapping data to obtain the weight value.

[0022] In some embodiments, the step of calculating a comprehensive score value for the target user according to the weight value includes:

[0023] Extract score values for the target growth data according to preset data dimensions to obtain target score values corresponding to each data dimension;

[0024] Perform maximum-minimum normalization processing on the target score values to obtain target scalar score values;

[0025] Perform weighted calculation on the target scalar score values and the weight values according to a preset formula to obtain the comprehensive score value.

[0026] In some embodiments, the step of constructing a target user portrait by performing portrait construction processing on the preset portrait attributes according to the weight value and the comprehensive score value includes:

[0027] Screen the candidate background data according to the target growth data to obtain target background data;

[0028] Perform layout processing on the preset portrait attributes according to the weight value and the comprehensive score value to obtain an initial user portrait;

[0029] Reconstruct the screen using the target background data and the initial user profile to obtain the target user profile.

[0030] In some embodiments, after the step of constructing a portrait for the preset portrait attributes according to the weight value and the comprehensive score value to obtain a target user portrait, the method includes:

[0031] Obtain a preset candidate layout interface;

[0032] Filter the candidate color data according to the target growth data to obtain target color data;

[0033] Reconstruct the candidate layout interface using the target color data and the target user profile to obtain a target layout interface;

[0034] Update the target layout interface according to preset first cycle parameters to display the target user profile.

[0035] In some embodiments, after the step of constructing a portrait for the preset portrait attributes according to the weight value and the comprehensive score value to obtain a target user portrait, the method includes:

[0036] Extract information from the target user profile according to preset second cycle parameters to obtain a current growth report;

[0037] Review the current growth report to obtain a review result;

[0038] Send the current growth report to a target platform and / or output an exception prompt message according to the review result.

[0039] To achieve the above object, a second aspect of the embodiments of the present application proposes a portrait construction device, the device includes:

[0040] A data acquisition module, configured to acquire target growth data, where the target growth data is data on the learning growth of a target user;

[0041] A data filtering module, configured to perform data filtering processing on the target growth data to obtain target data;

[0042] A prediction module, configured to perform prediction processing on the target data through a preset label prediction model to obtain target portrait labels of the target user;

[0043] A weight calculation module, configured to calculate weights for preset portrait attributes according to the target portrait labels to obtain a weight value corresponding to each preset portrait attribute;

[0044] A scoring calculation module, configured to calculate a growth score for the target user according to the weight value to obtain a comprehensive score value;

[0045] A portrait construction module, configured to perform portrait construction processing on the preset portrait attributes according to the weight value and the comprehensive score value to obtain a target user portrait.

[0046] To achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, which includes a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for realizing connection communication between the processor and the memory. When the program is executed by the processor, the method described in the first aspect above is realized.

[0047] To achieve the above object, a fourth aspect of the embodiments of the present application provides a storage medium, which is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to realize the method described in the first aspect above.

[0048] The portrait construction method, portrait construction device, electronic device and storage medium provided by the present application obtain target growth data, which is the learning growth data of the target user; and perform data filtering processing on the target growth data to obtain target data. Compared with the basic data in the related art, the acquisition source of the learning growth data of the target user is wider, which can meet the needs of multiple business scenarios, improve the applicability and comprehensiveness of the data, and can eliminate irrelevant data and abnormal data through data filtering processing, thereby improving the accuracy of the data. Further, through a preset label prediction model, the target data is predicted to obtain the target portrait label of the target user; the portrait of the target user can be intuitively defined through the portrait label. Finally, according to the target portrait label, the weight calculation of the preset portrait attributes is performed to obtain the weight value corresponding to each preset portrait attribute; according to the weight value, the growth score of the target user is calculated to obtain a comprehensive score value, which can count the proportion of different portrait attributes, clearly reflect the growth bias of the target user. Finally, according to the weight value and the comprehensive score value, portrait construction processing is performed on the preset portrait attributes to obtain a target user portrait, which can more clearly reflect the bias of the target user in different attributes, and can also reflect the comprehensive growth of the target user, making the target user portrait more referenceable and improving the accuracy of the constructed target user portrait. Description of the Drawings

[0049] Figure 1 is a flowchart of the portrait construction method provided by the embodiments of the present application;

[0050] Figure 2 is Figure 1 the flowchart of step S103 in

[0051] Figure 3 is Figure 1 the flowchart of step S104 in

[0052] Figure 4 is Figure 1 the flowchart of step S105 in

[0053] Figure 5 is Figure 1 the flowchart of step S106 in

[0054] Figure 6 is another flowchart of the image construction method provided by the embodiments of the present application;

[0055] Figure 7 is another flowchart of the image construction method provided by the embodiments of the present application;

[0056] Figure 8 is a schematic structural diagram of the image construction device provided by the embodiments of the present application;

[0057] Figure 9 is a schematic hardware structure diagram of the electronic device provided by the embodiments of the present application. Detailed implementation manners

[0058] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application.

[0059] It should be noted that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order from the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0061] First, several nouns involved in the present application are parsed:

[0062] Artificial Intelligence (AI): It is a new technical science that studies and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence; artificial intelligence is a branch of computer science. It attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a way similar to human intelligence. The research in this field includes robots, speech recognition, image recognition, natural language processing, and expert systems, etc. Artificial intelligence can simulate the information process of human consciousness and thinking. It is also a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0063] Natural Language Processing (NLP): NLP uses computers to process, understand, and apply human languages (such as Chinese, English, etc.). NLP is a branch of artificial intelligence and an interdisciplinary field of computer science and linguistics, and is often referred to as computational linguistics. Natural language processing includes syntactic analysis, semantic analysis, discourse understanding, etc. Natural language processing is commonly used in technical fields such as machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information intention recognition, information extraction and filtering, text classification and clustering, public opinion analysis, and opinion mining. It involves data mining related to language processing, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research, and linguistic research related to language computing.

[0064] Information Extraction (NER): A text processing technology that extracts factual information such as entities, relationships, events of a specified type from natural language texts and forms structured data output. Information extraction is a technology for extracting specific information from text data. Text data is composed of some specific units, such as sentences, paragraphs, and texts. Text information is exactly composed of some small specific units, such as words, phrases, sentences, paragraphs, or combinations of these specific units. Extracting noun phrases, personal names, place names, etc. from text data are all text information extraction. Of course, the information extracted by text information extraction technology can be various types of information.

[0065] User Profile: Also known as user persona, as an effective tool for depicting target users and connecting user demands with design directions, user profiles have been widely applied in various fields. In the actual operation process, we often connect the attributes, behaviors, and expected data of users with the simplest and most life-like words. As a virtual representative of actual users, the user personas formed by user profiles are not constructed outside of products and the market. The formed user personas need to be representative and can represent the main audience and target groups of the products.

[0066] Data cleaning: It refers to the last procedure of discovering and correcting identifiable errors in data files, including checking data consistency, handling invalid values, missing values, etc. Different from questionnaire review, data cleaning after input is generally completed by computers rather than manually.

[0067] Data deduplication: It means finding duplicate data in a data file set and deleting it, only keeping unique data units to eliminate redundant data. Data deduplication includes complete deduplication and incomplete deduplication. Complete deduplication means eliminating completely duplicate data, and completely duplicate data refers to data records with exactly the same field values in a data table. Incomplete deduplication means that in data cleaning, duplicate values with all field values being equal must be eliminated.

[0068] Clustering: It is to divide a data set into different classes or clusters according to a specific criterion (such as distance), so that the similarity of data objects within the same cluster is as large as possible, and at the same time, the difference of data objects not in the same cluster is also as large as possible. That is, after clustering, data of the same class is gathered together as much as possible, and data of different classes is separated as much as possible. Clustering is an unsupervised learning method.

[0069] UI interface: It is composed of multiple different basic visual elements. They form a complete interface effect through the combination of graphics, color matching, unity of materials and styles, and reasonable layout.

[0070] Currently, most portrait construction methods rely on the basic data of target users to construct user portraits, but the basic data often cannot cover multiple business scenarios, which often causes the problem of incomplete portraits, thus affecting the accuracy of portrait construction. Therefore, how to improve the accuracy of portrait construction has become a technical problem to be solved urgently.

[0071] Based on this, the embodiments of this application provide a portrait construction method, a portrait construction device, an electronic device, and a storage medium, aiming to improve the accuracy of portrait construction.

[0072] The image construction method, image construction device, electronic device, and storage medium provided by the embodiments of the present application will be specifically described through the following embodiments. First, the image construction method in the embodiments of the present application will be described.

[0073] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0074] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0075] The image construction method provided by the embodiments of the present application relates to the field of artificial intelligence technology. The image construction method provided by the embodiments of the present application can be applied to terminals, can also be applied to the server side, or can be software running on the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, can also be configured as a server cluster or distributed system composed of multiple physical servers, or can be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the image construction method, etc., but is not limited to the above forms.

[0076] This application can be used in numerous general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0077] Figure 1 is an alternative flowchart of the portrait construction method provided by an embodiment of this application. Figure 1 The method in may include but is not limited to steps S101 to S106.

[0078] Step S101, obtain target growth data, where the target growth data is the learning growth data of the target user;

[0079] Step S102, perform data filtering processing on the target growth data to obtain target data;

[0080] Step S103, perform prediction processing on the target data through a preset label prediction model to obtain the target portrait labels of the target user;

[0081] Step S104, calculate the weights of the preset portrait attributes according to the target portrait labels to obtain the weight values corresponding to each preset portrait attribute;

[0082] Step S105, calculate the growth score of the target user according to the weight values to obtain a comprehensive score value;

[0083] Step S106, perform portrait construction processing on the preset portrait attributes according to the weight values and the comprehensive score value to obtain the target user portrait.

[0084] Steps S101 to S106 illustrated in the embodiments of the present application, by obtaining target growth data, where the target growth data is the learning growth data of the target user; and performing data filtering processing on the target growth data to obtain target data. Compared with the basic data in the related art, the acquisition source of the learning growth data of the target user is wider, which can meet the requirements of various business scenarios, improve the applicability and comprehensiveness of the data, and the data filtering processing method can eliminate irrelevant data and abnormal data, thereby improving the accuracy of the data. Further, by using a preset label prediction model to perform prediction processing on the target data, target portrait labels of the target user are obtained; the portrait of the target user can be intuitively defined through the portrait labels. Finally, weight calculation is performed on the preset portrait attributes according to the target portrait labels to obtain the weight value corresponding to each preset portrait attribute; a growth score calculation is performed on the target user according to the weight value to obtain a comprehensive score value, which can count the proportion of different portrait attributes and clearly reflect the growth bias of the target user. Finally, portrait construction processing is performed on the preset portrait attributes according to the weight value and the comprehensive score value to obtain the target user portrait, which can more clearly reflect the bias of the target user in different attributes and also reflect the comprehensive growth situation of the target user, making the target user portrait more referenceable and improving the accuracy of the constructed target user portrait.

[0085] In step S101 of some embodiments, the target growth data can be obtained from different scenarios and different dimensions. Specifically, taking students as the target users as an example, data collection based on different scenarios can cover data from various business scenarios on campus, including but not limited to learning growth data such as campus activities, campus learning, campus entertainment, and campus interactions. For example, it includes learning growth data in aspects such as morality, intelligence, physical fitness, aesthetics, and labor. Data collection based on different dimensions can cover data from various platforms on campus, including but not limited to student classroom behavior data, student attendance data, student exam data, etc. on the academic affairs platform, smart campus platform, and teaching platform.

[0086] Further, in the specific data collection process, data collection can be carried out through various methods such as platform docking, background batch import, or upload using efficiency tools. This method can ensure that the collected data can cover multiple business scenarios, making the target growth data have good diversity and comprehensiveness.

[0087] It should be noted that in each specific embodiment of the present application, when it comes to relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of these data will comply with the relevant laws, regulations, and standards of the relevant countries and regions. In addition, when the embodiments of the present application need to obtain the user's target personal information, the user's separate permission or separate consent will be obtained through methods such as pop-up windows or redirecting to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0088] In step S102 of some embodiments, in order to improve the accuracy of the data, it is necessary to perform data filtering processing on the target growth data. Specifically, first, the target growth data is subjected to data cleaning through a regression function, that is, the target growth data is converted from a data form to an image form by using the regression function, and then the image is smoothed, and the target growth data is fitted to a multi-dimensional surface to eliminate noise data and obtain the initial data. Further, duplicate data is removed from the initial data, and the same initial data is removed to obtain the target data. By this means, irrelevant data and abnormal data can be removed, thereby improving the accuracy of the data.

[0089] Before step S103 of some embodiments, the portrait construction method further includes pre-training a label prediction model. The label prediction model can be constructed based on a BERT model and a convolutional neural network model. Among them, the label prediction model includes an embedding layer, a convolutional layer, an aggregation layer, and a prediction layer. The embedding layer is used to perform word embedding processing on the input data, convert the input data into a vector form, and obtain an embedded vector. The convolutional layer is used to extract features from the embedded vector, capture the entity features of the input data, and output entity features with higher importance. The clustering layer is used to aggregate entity features with higher relevance into the same class according to a preset clustering algorithm. The prediction layer is used to perform prediction processing on the entity features of different classes through a preset function to obtain the probability distribution of each class of entity features, and thus determine the target portrait label according to the probability distribution.

[0090] Please refer to Figure 2 , in some embodiments, the label prediction model includes an embedding layer, a convolutional layer, a clustering layer, and a prediction layer. Step S103 may include but is not limited to steps S201 to S205:

[0091] Step S201, perform word embedding processing on the target data through the embedding layer to obtain a target embedded vector;

[0092] Step S202, perform entity feature extraction on the target embedding vector through a convolutional layer to obtain target entity features;

[0093] Step S203, perform integration processing on the target entity features through a clustering layer to obtain central entity features;

[0094] Step S204, calculate the label probability of the central entity features through a preset function of the prediction layer to obtain a label probability value;

[0095] Step S205, determine the target portrait label according to the label probability value.

[0096] In step S201 of some embodiments, perform word embedding processing on the target data through an embedding layer, convert the target data from text form to vector form, and obtain a target embedding vector.

[0097] In step S202 of some embodiments, perform feature extraction on the target embedding vector through a convolutional layer, capture the entity features of the target embedding vector, and perform upsampling processing on the extracted entity features to obtain target entity features.

[0098] In step S203 of some embodiments, convert the target entity features into vector form through a clustering layer to obtain target entity vectors. Perform clustering processing on the target entity vectors based on the K-means algorithm. Specifically, select any number of target entity vectors as initial clustering centers, and pre-assign a corresponding class to each initial clustering center. Then calculate the Euclidean distance between each target entity vector serving as an initial clustering center and each initial clustering center. According to the distance value between the target entity vector and the initial clustering center, the target entity vector is classified into the class where the initial clustering center with the smallest distance value is located. In this way, multiple classes with the initial clustering center as the center point can be obtained. Further, according to the distribution of the target entity vectors in each class, re-determine the target clustering center of the class, that is, calculate the distance value between every two target entity vectors, and select the target entity vector that minimizes the sum of the distance values of the class as the target clustering center of the class, and use the target entity vector corresponding to this target clustering center as the central entity feature.

[0099] In step S204 of some embodiments, the preset function is the softmax function. Through the softmax function, a probability distribution of the central entity features belonging to each preset portrait label can be created, realizing the calculation of the label probability of the central entity features to obtain a label probability value. It should be noted that the label probability value here is represented in vector form, and the probability vector output by the softmax function is used as the label probability value corresponding to each preset portrait label.

[0100] In step S205 of some embodiments, since the magnitude of the probability value can represent the possibility of the target data belonging to each preset portrait label, the preset portrait label with the largest label probability value can be selected as the target portrait label.

[0101] Please refer to Figure 3 , in some embodiments, step S104 may include but is not limited to steps S301 to S304:

[0102] Step S301, obtain a preset relationship mapping table, which stores preset portrait labels and preset portrait attributes, and there is a mapping relationship between the preset portrait labels and the preset portrait attributes;

[0103] Step S302, extract the target portrait label from the preset portrait labels;

[0104] Step S303, extract the mapping data between the target portrait label and the preset portrait attributes from the relationship mapping table;

[0105] Step S304, perform a weight calculation on the preset portrait attributes according to the mapping data to obtain a weight value.

[0106] In step S301 of some embodiments, obtain a preset relationship mapping table, where the relationship mapping table stores preset portrait labels and preset portrait attributes, and there is a mapping relationship between the preset portrait labels and the preset portrait attributes. It should be noted that this mapping relationship can be in the form of one-to-one or one-to-many. For example, the preset portrait labels are age information, educational background information, learning and growth information, etc., and the preset portrait attributes include hobbies, health indicators, basic attributes, behavioral characteristics, and academic performance, etc.

[0107] In step S302 of some embodiments, since the target portrait label is one of the preset portrait labels, the same portrait label can be conveniently extracted from the preset portrait labels according to the label name or label value of the target portrait label.

[0108] In step S303 of some embodiments, after determining which label in the preset portrait labels is the target portrait label, the preset portrait attributes corresponding to this preset portrait label and the mapping relationship data between the two can be directly extracted from the relationship mapping table. For example, traverse the relationship mapping table according to the target portrait label to extract the mapping data between the target portrait label and the preset portrait attributes, where the mapping data includes the target portrait label, the preset portrait attributes, and the relevance between the target portrait label and the preset portrait attributes.

[0109] In step S304 of some embodiments, the relevance between the target portrait tags and the preset portrait attributes in the mapping data is extracted. By summing up this series of relevances, the total relevance is obtained, and each relevance is divided by the total relevance to obtain the proportion of each preset portrait attribute, and this ratio is used as the weight value of the preset portrait attribute.

[0110] For example, the target portrait tag corresponds to four preset portrait attributes (A, B, C, D), and the relevances between the target portrait tag and these four preset portrait attributes are 0.6, 0.4, 0.75, and 0.2 respectively. Then the total relevance is 0.6 + 0.4 + 0.75 + 0.2 = 1.95. Through division operation, the weight values of each preset portrait attribute are A = 0.6 / 1.95 = 0.31; B = 0.4 / 1.95 = 0.21; C = 0.75 / 1.95 = 0.38; D = 0.2 / 1.95 = 0.1.

[0111] Please refer to Figure 4 , in some embodiments, step S105 may include but is not limited to steps S401 to S403:

[0112] Step S401, extracting score values for the target growth data according to the preset data dimensions to obtain the target score value corresponding to each data dimension;

[0113] Step S402, performing maximum-minimum normalization processing on the target score value to obtain the target scalar score;

[0114] Step S403, performing weighted calculation on the target scalar score and the weight value according to the preset formula to obtain the comprehensive score value.

[0115] In step S401 of some embodiments, the preset data dimensions cover data from various campus platforms, including but not limited to the academic affairs platform, the smart campus platform, and the teaching platform. The target growth data includes student classroom behavior data, student attendance data, student exam data, etc. The target score value of the target growth data corresponding to each data dimension is extracted.

[0116] In step S402 of some embodiments, the target score value corresponding to each data dimension is quantified. Specifically, the target score value corresponding to each data dimension is converted to the same dimension. For example, the grade score, comprehensive performance score, etc. in the target score value are normalized to obtain the target scalar score, where the normalization processing can be maximum-minimum normalization processing. Specifically, the method of linearizing the original feature value data of the above data dimensions is converted to the range of [Xmin..Xmax], and the formula is shown as formula (1):

[0117]

[0118] In this way, the proportional scaling of the target growth data can be achieved. Among them, Xnorm is the normalized target scalar score, X is the target score value, and Xmax and Xmin are the maximum and minimum values of the target score value respectively.

[0119] In step S403 of some embodiments, the preset formula can be expressed as shown in formula (2). Through the preset formula, the comprehensive score value y can be conveniently calculated.

[0120]

[0121] Among them, x1, x2,..., xn are the target scalar scores of each data dimension, and w1, w2,..., wn are the weight values of each preset portrait attribute.

[0122] Through the above-mentioned target score value and comprehensive score value, the overall score situation of the target object can be more conveniently analyzed. The scores at different stages and different dimensions are compared, and the weak points of the target user are analyzed. Thus, the proportion of different portrait attributes is statistically calculated, clearly reflecting the growth bias of the target user.

[0123] Please refer to Figure 5 , in some embodiments, step S106 may include but is not limited to steps S501 to S503:

[0124] Step S501, screening the candidate background data according to the target growth data to obtain the target background data;

[0125] Step S502, arranging the preset portrait attributes according to the weight value and the comprehensive score value to obtain the initial user portrait;

[0126] Step S503, reconstructing the picture through the target background data and the initial user portrait to obtain the target user portrait.

[0127] In step S501 of some embodiments, the background keywords corresponding to the target user are extracted from the target growth data, and the candidate background data is screened according to the background keywords. The background features containing the background keywords are selected, and this background feature is used as the target background data. Taking a student as the target user as an example, the target growth data contains data such as the school, grade, and class where the student is located. Based on the school, grade, and class where the student is located, the corresponding background features are selected from the candidate background data. The background feature can be a certain solid color feature (such as red, etc.), and this background feature is used as the target background data. This target background data can be used as the background image of the target user portrait.

[0128] In step S502 of some embodiments, different preset portrait attributes are sorted in descending order according to the magnitude of the weight values to obtain a portrait attribute sequence. The preset portrait attributes are sequentially filled into a preset layout panel according to the portrait attribute sequence. The layout panel can reflect the proportion of different preset portrait attributes. For example, the layout panel can be in the form of a circular block, a six-dimensional graph, a bar graph, etc., and the distribution and proportion of different preset portrait attributes are reflected through the layout panel to obtain an initial user portrait.

[0129] In step S503 of some embodiments, pixel features in the target background data are extracted. This pixel feature is used as the background image, and the initial user portrait is used as the foreground image. The background image and the foreground image are fused to achieve image reconstruction and obtain the target user portrait. The target user portrait can reflect various information such as the basic attributes, behavioral characteristics, hobbies, and nutritional characteristics of the target user. The target user portrait can more clearly reflect the preference of the target user in different attributes and can also reflect the comprehensive growth of the target user, improving the accuracy of portrait construction.

[0130] Please refer to Figure 6 , in some embodiments, after step S106, the portrait construction method may include but is not limited to steps S601 to S604:

[0131] Step S601, obtain a preset candidate layout interface;

[0132] Step S602, screen the candidate color data according to the target growth data to obtain the target color data;

[0133] Step S603, perform image reconstruction on the candidate layout interface through the target color data and the target user portrait to obtain the target layout interface;

[0134] Step S604, update the target layout interface according to the preset first cycle parameter to display the target user portrait.

[0135] In step S601 of some embodiments, a preset candidate layout interface is obtained from the layout interface template library. The candidate layout interface is a UI interface, and this UI interface can be used for the display of images, texts, etc. It makes texts, images, etc. visible for the target object to observe. The target object here can be an observer, a target user, or other personnel, without limitation.

[0136] In step S602 of some embodiments, color keywords corresponding to the target user are extracted from the target growth data. The candidate color data is screened according to the color keywords, and color features containing the color keywords are selected. This color feature is used as the background color of the layout interface. Taking students as the target users as an example, the target growth data includes data such as the school, grade, and class where the student is located, the type of the school, and the school logo. Based on the school, grade, and class where the student is located, the corresponding color feature is selected from the candidate color data. This color feature can be a certain pure color feature (such as red, etc.) or a color combination feature. This color feature is used as the background color of the layout interface to obtain the target color data.

[0137] In step S603 of some embodiments, the pixel features in the target color data are extracted. This pixel feature is used as the background color of the layout interface, and the target user portrait is used as the foreground image. The background color of the layout interface and the target user portrait are fused to achieve image reconstruction, and the target layout interface is obtained.

[0138] In step S604 of some embodiments, the first cycle parameter can be set according to actual business needs without limitation. To improve the display efficiency of the target user portrait within a fixed time, the first cycle parameter can be set to 5 seconds, that is, the target layout interface is updated every 5 seconds to display the target user portraits of different target users, and the observation efficiency of the target user portraits of the target users is achieved.

[0139] Please refer to Figure 7 , in some embodiments, after step S106, the portrait construction method may further include but is not limited to steps S701 to S703:

[0140] Step S701: Extract information from the target user portrait according to the preset second cycle parameter to obtain the current growth report;

[0141] Step S702: Perform a review process on the current growth report to obtain a review result;

[0142] Step S703: Send the current growth report to the target platform and / or output an exception prompt message according to the review result.

[0143] In step S701 of some embodiments, the second cycle parameter can be set according to actual business needs without limitation. To better monitor the changes in the target user portrait of the target user, the second cycle parameter can be set to one week or one month. For example, information is extracted from the target user portrait every week, the current growth data of the target user is obtained, and the current growth data is input into a preset report template to obtain the current growth report.

[0144] In step S702 of some embodiments, the current growth report can be reviewed manually, and according to the preset verification rules, it is determined whether there are abnormalities in the data of the current growth report, so as to generate a review result. In addition, the growth report of the previous stage can also be compared and analyzed with the current growth report to obtain the deviation of the growth data, and the review result can be generated according to the deviation of the growth data. In other embodiments, other methods can also be used to review the current growth report to obtain a review result, which is not limited to this.

[0145] In step S703 of some embodiments, if the review result is normal, the current growth report is sent to the target platform. The target platform can be the current mainstream platform (such as smart campus, WeChat, official account, etc.), or other fixed third parties, so that different personnel can view the current growth report of the target object from the target platform; if the review result is abnormal, an abnormal prompt message is output to the administrator, so that the administrator can locate and troubleshoot the abnormal data, and timely correct the data of the current growth report, thereby improving the efficiency of data processing.

[0146] The portrait construction method of the embodiment of the present application obtains target growth data, where the target growth data is the learning growth data of the target user; and performs data filtering processing on the target growth data to obtain target data. Compared with the basic data in the related art, the acquisition source of the learning growth data of the target user is wider, which can meet the needs of various business scenarios, improve the applicability and comprehensiveness of the data, and can eliminate irrelevant data and abnormal data through data filtering processing, thereby improving the accuracy of the data. Further, the target data is predicted by a preset label prediction model to obtain the target portrait label of the target user; the portrait of the target user can be defined intuitively through the portrait label. Finally, the weight of each preset portrait attribute is calculated according to the target portrait label to obtain the weight value corresponding to each preset portrait attribute; the growth score of the target user is calculated according to the weight value to obtain the comprehensive score value, which can count the proportion of different portrait attributes and clearly reflect the growth bias of the target user. Finally, the preset portrait attributes are processed for portrait construction according to the weight value and the comprehensive score value to obtain the target user portrait, which can more clearly reflect the bias of the target user in different attributes and can also reflect the comprehensive growth of the target user, making the target user portrait more referenceable and improving the accuracy of the constructed target user portrait.

[0147] Please refer to Figure 8 , the embodiment of the present application also provides a portrait construction device, which can implement the above portrait construction method. The device includes:

[0148] The data acquisition module 801 is used to acquire target growth data, where the target growth data is the learning growth data of the target user;

[0149] The data filtering module 802 is used to perform data filtering processing on the target growth data to obtain target data;

[0150] The prediction module 803 is used to perform prediction processing on the target data through a preset label prediction model to obtain the target portrait labels of the target user;

[0151] The weight calculation module 804 is used to calculate the weights of preset portrait attributes according to the target portrait labels to obtain the weight values corresponding to each preset portrait attribute;

[0152] The score calculation module 805 is used to calculate the growth score of the target user according to the weight values to obtain a comprehensive score value;

[0153] The portrait construction module 806 is used to perform portrait construction processing on the preset portrait attributes according to the weight values and the comprehensive score value to obtain the target user portrait.

[0154] In some embodiments, the label prediction model includes an embedding layer, a convolutional layer, a clustering layer, and a prediction layer. The prediction module 803 includes:

[0155] An embedding unit, which is used to perform word embedding processing on the target data through the embedding layer to obtain a target embedding vector;

[0156] An entity extraction unit, which is used to perform entity feature extraction on the target embedding vector through the convolutional layer to obtain target entity features;

[0157] An integration unit, which is used to perform integration processing on the target entity features through the clustering layer to obtain central entity features;

[0158] A probability calculation unit, which is used to perform label probability calculation on the central entity features through a preset function of the prediction layer to obtain label probability values;

[0159] A label determination unit, which is used to determine the target portrait labels according to the label probability values.

[0160] In some embodiments, the weight calculation module 804 includes:

[0161] A table acquisition unit, which is used to acquire a preset relationship mapping table. The relationship mapping table stores preset portrait labels and preset portrait attributes, and there is a mapping relationship between the preset portrait labels and the preset portrait attributes;

[0162] A label extraction unit, which is used to extract the target portrait labels from the preset portrait labels;

[0163] A mapping data extraction unit for extracting mapping data between target portrait tags and preset portrait attributes from a relational mapping table;

[0164] A calculation unit for calculating weights of the preset portrait attributes according to the mapping data to obtain weight values.

[0165] In some embodiments, the scoring calculation module 805 includes:

[0166] A score extraction unit for extracting score values of the target growth data according to preset data dimensions to obtain target score values corresponding to each data dimension;

[0167] A normalization unit for performing maximum-minimum normalization processing on the target score values to obtain target scalar scores;

[0168] A weighted calculation unit for performing weighted calculation on the target scalar scores and the weight values according to a preset formula to obtain a comprehensive score value.

[0169] In some embodiments, the portrait construction module 806 includes:

[0170] A background screening unit for screening candidate background data according to the target growth data to obtain target background data;

[0171] A portrait layout unit for performing layout processing on the preset portrait attributes according to the weight values and the comprehensive score values to obtain an initial user portrait;

[0172] A portrait reconstruction unit for reconstructing the screen through the target background data and the initial user portrait to obtain a target user portrait.

[0173] In some embodiments, the portrait construction device further includes a display module, specifically including:

[0174] An interface acquisition unit for acquiring a preset candidate layout interface;

[0175] A color screening unit for screening candidate color data according to the target growth data to obtain target color data;

[0176] An interface reconstruction unit for reconstructing the screen of the candidate layout interface through the target color data and the target user portrait to obtain a target layout interface;

[0177] A display unit for updating the target layout interface according to preset first cycle parameters to display the target user portrait.

[0178] In some embodiments, the portrait construction device further includes an output module, specifically including:

[0179] An information extraction unit, configured to extract information from the target user portrait according to a preset second cycle parameter to obtain a current growth report;

[0180] A review unit, configured to perform a review process on the current growth report to obtain a review result;

[0181] An output unit, configured to send the current growth report to the target platform and / or output an exception prompt message according to the review result.

[0182] The specific implementation manner of this portrait construction device is basically the same as the specific embodiments of the above portrait construction method, and will not be elaborated here.

[0183] An embodiment of the present application further provides an electronic device, which includes: a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for implementing connection communication between the processor and the memory. When the program is executed by the processor, the above portrait construction method is implemented. The electronic device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0184] Please refer to Figure 9 , Figure 9 , which schematically shows the hardware structure of an electronic device in another embodiment. The electronic device includes:

[0185] A processor 901, which can be implemented in a general-purpose CPU (Central Processing Unit, central processor), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., and is configured to execute relevant programs to implement the technical solutions provided by the embodiments of the present application;

[0186] A memory 902, which can be implemented in the form of a read-only memory (ReadOnlyMemory, ROM), a static storage device, a dynamic storage device, or a random access memory (RandomAccessMemory, RAM), etc. The memory 902 can store an operating system and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 902, and the processor 901 is called to execute the portrait construction method of the embodiments of the present application;

[0187] An input / output interface 903, configured to implement information input and output;

[0188] A communication interface 904, which is used to implement the communication interaction between this device and other devices. It can achieve communication through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0189] A bus 905, which transmits information between various components of the device (such as a processor 901, a memory 902, an input / output interface 903, and a communication interface 904);

[0190] Among them, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904 achieve communication connections with each other inside the device through the bus 905.

[0191] The embodiment of this application also provides a storage medium. The storage medium is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the above portrait construction method.

[0192] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory optionally includes a memory remotely set relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0193] The image construction method, image construction device, electronic device, and storage medium provided by the embodiments of the present application obtain target growth data, which is the learning growth data of a target user; and perform data filtering processing on the target growth data to obtain target data. Compared with the basic data in the related art, the acquisition source of the learning growth data of the target user is wider, which can meet the requirements of various business scenarios, improve the applicability and comprehensiveness of the data, and can eliminate irrelevant data and abnormal data through data filtering processing, thereby improving the accuracy of the data. Further, through a preset label prediction model, the target data is predicted to obtain the target image label of the target user; the image of the target user can be intuitively defined through the image label. Finally, weight calculation is performed on the preset image attributes according to the target image label to obtain the weight value corresponding to each preset image attribute; a growth score calculation is performed on the target user according to the weight value to obtain a comprehensive score value, which can count the proportion of different image attributes and clearly reflect the growth bias of the target user. Finally, according to the weight value and the comprehensive score value, image construction processing is performed on the preset image attributes to obtain the target user image, which can more clearly reflect the bias of the target user in different attributes and can also reflect the comprehensive growth situation of the target user, making the target user image more referenceable and improving the accuracy of the constructed target user image.

[0194] The embodiments described in the embodiments of the present application are for more clearly explaining the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0195] Those skilled in the art can understand that Figure 1-7 the technical solutions shown in do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown, or combine certain steps, or different steps.

[0196] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0197] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices can be implemented as software, firmware, hardware, and their appropriate combinations.

[0198] In the description of the present application and the above-mentioned drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0199] It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships can exist. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally means that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or its similar expression means any combination of these items, including any combination of single items (ones) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0200] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the above-mentioned division of units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces, and the indirect coupling or communication connection of devices or units can be in electrical, mechanical, or other forms.

[0201] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0202] In addition, in each embodiment of the present application, each functional unit can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0203] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The aforementioned storage medium includes: various media that can store programs, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0204] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings, and thus do not limit the scope of rights of the embodiments of the present application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of rights of the embodiments of the present application.

Claims

1. An image construction method, characterized in that, the method includes: Obtain target growth data, where the target growth data is the learning growth data of a target user; Perform data filtering processing on the target growth data to obtain target data; Perform prediction processing on the target data through a preset label prediction model to obtain the target portrait labels of the target user; Calculate the weights of preset portrait attributes according to the target portrait labels to obtain the weight values corresponding to each preset portrait attribute; Calculate the growth score of the target user according to the weight values to obtain a comprehensive score value; Perform portrait construction processing on the preset portrait attributes according to the weight values and the comprehensive score value to obtain a target user portrait; The step of calculating the weights of preset portrait attributes according to the target portrait labels to obtain the weight values corresponding to each preset portrait attribute includes: Obtain a preset relationship mapping table, where the relationship mapping table stores preset portrait labels and preset portrait attributes, and there is a mapping relationship between the preset portrait labels and the preset portrait attributes; extract the target portrait labels from the preset portrait labels; extract the mapping data between the target portrait labels and the preset portrait attributes from the relationship mapping table; calculate the weights of the preset portrait attributes according to the mapping data to obtain the weight values; The step of calculating the growth score of the target user according to the weight values to obtain a comprehensive score value includes: Extract score values of the target growth data according to preset data dimensions to obtain target score values corresponding to each data dimension; perform maximum-minimum normalization processing on the target score values to obtain target scalar score values; perform weighted calculation on the target scalar score values and the weight values according to a preset formula to obtain the comprehensive score value; The step of performing portrait construction processing on the preset portrait attributes according to the weight values and the comprehensive score value to obtain a target user portrait includes: Screen candidate background data according to the target growth data to obtain target background data; perform layout processing on the preset portrait attributes according to the weight values and the comprehensive score value to obtain an initial user portrait; Extract pixel features in the target background data as the background image, use the initial user portrait as the foreground image, and perform image fusion on the background image and the foreground image to realize image reconstruction and obtain the target user portrait.

2. The image construction method according to claim 1, characterized in that, the label prediction model includes an embedding layer, a convolutional layer, a clustering layer, and a prediction layer. The step of performing prediction processing on the target data through a preset label prediction model to obtain the target portrait labels of the target user includes: Perform word embedding processing on the target data through the embedding layer to obtain a target embedding vector; Extract entity features of the target embedding vector through the convolutional layer to obtain target entity features; Perform integration processing on the target entity features through the clustering layer to obtain central entity features; Calculate the label probability of the central entity feature through a preset function of the prediction layer to obtain a label probability value; Determine the target portrait label according to the label probability value.

3. The portrait construction method according to any one of claims 1 to 2, characterized in that, after the step of performing portrait construction processing on the preset portrait attributes according to the weight value and the comprehensive score value to obtain a target user portrait, the method includes: Obtain a preset candidate layout interface; Screen the candidate color data according to the target growth data to obtain target color data; Reconstruct the candidate layout interface through the target color data and the target user portrait to obtain a target layout interface; Update the target layout interface according to a preset first cycle parameter to display the target user portrait.

4. The portrait construction method according to any one of claims 1 to 2, characterized in that, after the step of performing portrait construction processing on the preset portrait attributes according to the weight value and the comprehensive score value to obtain a target user portrait, the method includes: Extract information from the target user portrait according to a preset second cycle parameter to obtain a current growth report; Perform a review process on the current growth report to obtain a review result; Send the current growth report to a target platform and / or output an exception prompt message according to the review result.

5. A portrait construction device, characterized in that, the device includes: A data acquisition module for acquiring target growth data, where the target growth data is data of the learning growth of a target user; A data filtering module for performing data filtering processing on the target growth data to obtain target data; A prediction module for performing prediction processing on the target data through a preset label prediction model to obtain a target portrait label of the target user; A weight calculation module for calculating weights for preset portrait attributes according to the target portrait label to obtain a weight value corresponding to each preset portrait attribute; A score calculation module for calculating a growth score for the target user according to the weight value to obtain a comprehensive score value; A portrait construction module for performing portrait construction processing on the preset portrait attributes according to the weight value and the comprehensive score value to obtain a target user portrait; The calculating weights for preset portrait attributes according to the target portrait label to obtain a weight value corresponding to each preset portrait attribute includes: Obtain a preset relationship mapping table, where the relationship mapping table stores preset portrait labels and preset portrait attributes, and there is a mapping relationship between the preset portrait labels and the preset portrait attributes; extract the target portrait label from the preset portrait labels; extract mapping data between the target portrait label and the preset portrait attributes from the relationship mapping table; calculate weights for the preset portrait attributes according to the mapping data to obtain the weight value; The calculating a growth score for the target user according to the weight value to obtain a comprehensive score value includes: Extract score values from the target growth data according to preset data dimensions to obtain the target score value corresponding to each data dimension; perform maximum-minimum normalization processing on the target score value to obtain the target scalar score; perform weighted calculation on the target scalar score and the weight value according to a preset formula to obtain the comprehensive score value; Perform portrait construction processing on the preset portrait attributes according to the weight value and the comprehensive score value to obtain a target user portrait, including: Screen the candidate background data according to the target growth data to obtain the target background data; perform layout processing on the preset portrait attributes according to the weight value and the comprehensive score value to obtain the initial user portrait; Extract the pixel features in the target background data as the background image, use the initial user portrait as the foreground image, and perform image fusion on the background image and the foreground image to realize picture reconstruction and obtain the target user portrait.

6. An electronic device, Characterized in that, The electronic device includes a memory, a processor, a program stored on the memory and executable on the processor, and a data bus for realizing connection communication between the processor and the memory. When the program is executed by the processor, the steps of the portrait construction method according to any one of claims 1 to 4 are realized.

7. A storage medium, the storage medium is a computer-readable storage medium for computer-readable storage, Characterized in that, The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to realize the steps of the portrait construction method according to any one of claims 1 to 4.

Images