WO2023070844A1

WO2023070844A1 - Intelligent design method for personalized discount coupon, and electronic apparatus and storage medium

Info

Publication number: WO2023070844A1
Application number: PCT/CN2021/136416
Authority: WO
Inventors: 苏新铎; 王晓亮; 陈�光
Original assignee: 广州广电运通金融电子股份有限公司
Priority date: 2021-10-25
Filing date: 2021-12-08
Publication date: 2023-05-04
Also published as: CN113919882A

Abstract

An intelligent design method for a personalized discount coupon, and an electronic apparatus and a storage medium. The method comprises: acquiring a video stream that is captured by a camera (S101); executing a target tracking task on a pedestrian and a commodity in the video stream, so as to obtain a tracking result (S102); calculating commodity sales condition features according to the tracking result, wherein the commodity sales condition features at least comprise the number of times the commodity has been viewed, the number of times the commodity has been inquired about, the number of times the commodity has been taken, and the number of times the commodity has been placed into a shopping cart (S103); training a regression model by using the commodity sales condition features, so as to obtain a sales prediction model (S104); predicting, by using the sales prediction model, the sales volume of the commodity under a given discount (S105); and on the basis of a predicted sales volume, generating a discount coupon under the condition of profit maximization (S106). The method can assist a merchant in rationally controlling marketing costs and can also increase profits, without violating the privacy of users, such that the problems of high marketing costs, a low sales volume and low profits caused by a previous method for issuing discount coupons according to personal experience are solved.

Description

Intelligent design method, electronic device and storage medium for personalized discount coupon

technical field

The present application relates to the field of computer technology, in particular to an intelligent design method, electronic device and storage medium of a personalized discount coupon.

Background technique

Issuing discount coupons is a common promotion behavior, and shopping malls use discount coupons to attract users to consume. At present, for large shopping malls or supermarkets, merchants rely heavily on personal experience for discount coupons, and cannot intelligently generate discount coupon amounts for specific or different products, resulting in increased marketing costs, but they cannot increase product sales and profits.

At present, no effective solution has been proposed for the problems of increased marketing costs, low sales volume and low profits caused by the way of distributing discount coupons through personal experience in related technologies.

Contents of the invention

The embodiment of the present application provides an intelligent design method, electronic device, and storage medium for personalized discount coupons, so as to at least solve the problems of increased marketing costs, low sales volume and low profits in the related art by issuing discount coupons based on personal experience question.

In the first aspect, the embodiment of the present application provides an intelligent design method for a personalized discount coupon, the method includes the following steps:

Obtain the video stream captured by the camera;

Perform target tracking tasks on pedestrians and commodities in the video stream to obtain tracking results;

According to the tracking results, the product sales status feature is calculated, and the product sales status feature at least includes the number of times the product is viewed, the number of times the product is consulted, the number of times the product is taken, and the number of times the product is added to the shopping cart;

Utilize the characteristic training regression model of described commodity sales situation, obtain sales volume prediction model;

Using the sales forecast model to predict the sales volume of the commodity under a given discount;

According to the predicted sales volume, discount coupons are generated under the condition of maximizing profit.

In some of these embodiments, the product sales status feature also includes the number of times the product has been browsed for more than a preset threshold but not taken, and the number of times the product has been taken but not added to the shopping cart.

In some of these embodiments, after the acquisition of the video stream captured by the camera, the method further includes:

Identify the identity of the pedestrian through the image classification model, and judge whether the identity of the pedestrian is a service person or a customer;

When the identity of a pedestrian is identified as a customer, generate the temporary ID of the customer, and perform the pedestrian attribute identification task on the customer, and add the extracted customer's age interval feature and gender feature to the commodity sales status feature, when the When the customer leaves the shooting area, the temporary ID is deleted.

In some of these embodiments, before identifying the identity of the pedestrian through the image classification model and judging whether the identity of the pedestrian is a service person or a customer, the method further includes:

Detecting the rectangular area of the pedestrian position appearing in the video stream picture through the target detection model, and outputting the target detection result;

cutting the pedestrian image according to the target detection result, and performing quality evaluation on the cut pedestrian image through the quality evaluation module;

If the image of the pedestrian does not conform to the preset quality evaluation standard of the quality evaluation module, the identity information of the pedestrian is not identified.

In some of these embodiments, after calculating the product sales status characteristics according to the tracking results, the method further includes:

For a new product without the characteristics of the sales status of the product, the most similar product with the characteristics of the sales status of the product is found through the inherent attributes of the new product.

Calculating the commodity attention score according to the characteristics of the commodity sales status;

Products are screened according to the attention score and sales volume of the products.

In some of these embodiments, the regression model is a linear regression model, a random forest model or a gradient boosting decision tree model.

In some of these embodiments, the generating discount coupons under the condition of maximizing profits includes:

Calculation of the discount value at which the profit is maximized;

Call the preset discount coupon template, which includes the page and appearance of the discount coupon;

Write the discount coupon attribute, discount date and the discount value into the corresponding position of the discount coupon template, generate the discount coupon, and issue the discount coupon to the customer through the membership system.

In the second aspect, the embodiment of the present application provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to implement the above-mentioned first aspect when running the computer program. The intelligent design method of the personalized discount coupon described above.

In a third aspect, an embodiment of the present application provides a storage medium, in which a computer program is stored, wherein the computer program is configured to implement the personalized discount coupon as described in the first aspect when running. Smart design approach.

Compared with related technologies, the embodiment of the present application provides an intelligent design method for personalized discount coupons, which obtains video streams captured by cameras, performs target tracking tasks on pedestrians and commodities in the video streams, and obtains tracking results; Streaming data is processed to realize the analysis of customer (pedestrian) behavior, and according to the tracking results, the product sales status feature is calculated, and the regression model is trained using the product sales status feature to obtain a sales forecast model, without using the customer's personal information. In the case of violating customer privacy, analyze the sales status of the product, and accurately predict the sales of the product under different discount amounts through the sales forecast model, and realize the detection of the sales status of the product in a single network (sales forecast model) , to reduce the reasoning time by sharing most of the calculations, and achieve the purpose of performing reasoning and prediction at the video frame rate, using the sales prediction model to predict the sales of products under a given discount, so that it is convenient for a large number of different products to be quickly predicted Sales volume is highly intelligent and practical. According to the forecasted sales volume, discount coupons are generated under the condition of maximizing profits. In this way, intelligently designing discount coupons without infringing on user privacy can not only help merchants reasonably control marketing Costs can also increase profits, which solves the problems of increased marketing costs and low sales and profits caused by the previous way of issuing discount coupons through personal experience.

Description of drawings

The drawings described here are used to provide a further understanding of the application and constitute a part of the application. The schematic embodiments and descriptions of the application are used to explain the application and do not constitute an improper limitation to the application. In the attached picture:

Fig. 1 is a schematic flow chart of a first intelligent design method of a personalized discount coupon according to an embodiment of the present application;

Fig. 2 is a second schematic flow chart of an intelligent design method for a personalized discount coupon according to an embodiment of the present application;

Fig. 3 is a third schematic flow chart of an intelligent design method for a personalized discount coupon according to an embodiment of the present application;

FIG. 4 is a schematic flow diagram of obtaining the number of times a commodity has been viewed according to Embodiment 1 of the present application;

FIG. 5 is a schematic flow diagram of obtaining the number of consultations or consultation time of a product according to Embodiment 1 of the present application;

FIG. 6 is a schematic flow diagram of obtaining the number of times a product is taken and the number of times a product is added to a shopping cart according to an embodiment of the present application;

Fig. 7 is a fourth schematic flowchart of an intelligent design method for a personalized discount coupon according to an embodiment of the present application;

Fig. 8 is a schematic diagram of a product discount and sales song according to an embodiment of the present application;

FIG. 9 is a schematic flow chart of generating discount coupons under a profit maximization condition according to an embodiment of the present application;

FIG. 10 is a schematic diagram of an internal structure of an electronic device according to an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application. In addition, it can also be understood that although such development efforts may be complex and lengthy, for those of ordinary skill in the art relevant to the content disclosed in this application, the technology disclosed in this application Some design, manufacturing or production changes based on the content are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

Reference in this application to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The occurrences of this phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is understood explicitly and implicitly by those of ordinary skill in the art that the embodiments described in this application can be combined with other embodiments without conflict.

Unless otherwise defined, the technical terms or scientific terms involved in the application shall have the usual meanings understood by those with ordinary skill in the technical field to which the application belongs. Words such as "a", "an", "an" and "the" involved in this application do not indicate a limitation on quantity, and may indicate singular or plural numbers. The terms "comprising", "comprising", "having" and any variations thereof involved in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product or process that includes a series of steps or modules (units). The apparatus is not limited to the listed steps or units, but may further include steps or units not listed, or may further include other steps or units inherent to the process, method, product or apparatus. The words "connected", "connected", "coupled" and similar words mentioned in this application are not limited to physical or mechanical connection, but may include electrical connection, no matter it is direct or indirect. "Multiple" referred to in the present application means greater than or equal to two. "And/or" describes the association relationship of associated objects, indicating that there may be three types of relationships. For example, "A and/or B" may indicate: A exists alone, A and B exist simultaneously, and B exists independently. The terms "first", "second", and "third" involved in this application are only used to distinguish similar objects, and do not represent a specific ordering of objects.

The invention proposes an intelligent design method for personalized discount coupons.

Fig. 1 is a schematic flow chart of a first intelligent design method of a personalized discount coupon according to an embodiment of the present application. With reference to Fig. 1, in an embodiment of the present invention, the intelligent design method of a personalized discount coupon proposed by the present invention is applied In a large shopping mall/supermarket with a camera installed in the commodity area, the method includes the following steps:

Step S101, obtaining the video stream captured by the camera;

Step S102, perform target tracking tasks on pedestrians and commodities in the video stream to obtain tracking results; for example, process the video stream data to realize the analysis of customer (pedestrian) behavior;

It should be noted that the target tracking task in this embodiment is a single-step multi-target tracking task, product detection and Re-ID (pedestrian re-identification) feature extraction can be realized simultaneously through a single model, in a single network (sales forecast model ) to complete product detection and identity embedding at the same time, so as to reduce the reasoning time by sharing most of the calculations, and achieve the purpose of performing reasoning and prediction at the video frame rate; in this embodiment, the target tracking task uses FairMOT (FairMOT, the full name is Fair Multi -Object Tracking) model to achieve, of course, in some other embodiments, other models can also be used to achieve, not specifically limited here;

Step S103, calculate the product sales status feature according to the tracking results, the product sales status feature at least includes the number of times the product is viewed, the number of times the product is consulted, the number of times the product is picked up, and the number of times the product is added to the shopping cart; of course, in some other embodiments , the product sales status feature may also include the basic information feature of the product or other features, and no specific features are included here; for example, the basic information feature includes information such as product identification, product name, product unit price, and product discount;

It is easy to understand that the more characteristics of commodity sales status, the better the prediction effect of the subsequent sales forecast model and the higher the prediction accuracy; As a result, the product sales status features are calculated. In other words, the product sales status features include at least the number of times the product was viewed in the previous time span, the number of times the product was consulted in the previous time span, and the number of times the product was taken in the previous time span. The number of fetch times and the number of times the products in the previous time span were added to the shopping cart, where the last time span refers to the time period during which the price and discount of the last product continued to remain unchanged, such as the time of the last promotional period of the product: that is, 1 From 0:00:00 on January 6th to 0:00:00 on January 9th, during this period, for example, the price of the product is 200 yuan, and the discount amount is 180 yuan (10% off); Time: from 0:00:00 on January 15th to 0:00:00 on January 20th. During this period, for example, the price of the product is 100 yuan, and the discount amount is 80 yuan (20% off); in addition, the Those skilled in the art can also extract the characteristics of the sales status of the commodity through the existing video algorithm, which will not be repeated here.

In step S104, the regression model is trained using the characteristics of commodity sales conditions to obtain a sales forecast model; in this way, the sales forecast is realized;

Step S105, use the sales forecast model to predict the sales volume of the product under a given discount; in this way, it is convenient to quickly predict the sales volume of a large number of different products, with high intelligence and strong practicability; wherein, according to user needs, the given discount can be Set according to user needs, and the given discount can be a given amount or a given discount (for example, 30% off, 20% off, 10% off, etc.), which is not specifically limited here;

Step S106, according to the predicted sales volume, a discount coupon is generated under the condition of maximizing profit. In this way, it can not only help merchants to reasonably control marketing costs, but also increase profits. It solves the problems of increasing marketing costs and low sales and profits caused by issuing discount coupons through personal experience. It should be noted that this application can generate different discount coupons for different commodities, and the discount coupons do not distinguish between customers;

It should be noted that the above steps S101 to S106 in this embodiment are all completed without infringing on user privacy, wherein not infringing on user privacy refers to not using personal private data of customers in this method. It is worth noting that, in the scenario of the embodiment of the present application, the above steps can be used to intelligently generate discount coupons for all commodities in the mall.

Through the above steps S101 to S106, the video stream captured by the camera is obtained, and the target tracking task is performed on pedestrians and commodities in the video stream to obtain the tracking result; thus, the video stream data is processed to realize the analysis of customer (pedestrian) behavior, According to the tracking results, the product sales status characteristics are calculated, and the regression model is trained using the product sales status features to obtain a sales forecast model. Without using the customer's personal information, the product sales status is analyzed without infringing on the customer's privacy, and passed The sales forecast model can accurately predict the sales of products under different discount amounts, and realize the detection of product sales in a single network (sales forecast model), so as to reduce the reasoning time by sharing most of the calculations, and realize the video The frame rate performs reasoning and forecasting purposes, using the sales forecasting model to predict the sales of products under a given discount, so that it is convenient to quickly predict the sales of a large number of different products, with high intelligence and strong practicability. According to the predicted sales , to generate discount coupons under the condition of profit maximization. In this way, without infringing on user privacy, intelligently designing discount coupons can not only help merchants reasonably control marketing costs but also increase profits. The method of discount coupons has caused the problems of increased marketing costs, low sales and low profits.

Because there are still a large number of potential customers in the shopping process, many potential customers are in a wait-and-see state. Once they get good news, they can easily be converted into new customers. The number of times it was not picked up and the number of times the item was picked up but not added to the shopping cart. In this way, by increasing the characteristics of a large number of potential customers' behaviors on commodities, it is convenient to better train the sales prediction model and improve the prediction accuracy; wherein, the preset threshold is set according to user needs. For example, the preset threshold can be 1 day, 2 days or more, not specifically limited here.

Further, in some embodiments, the product sales status is characterized by the number of times the product is browsed for more than a preset threshold but not taken, the number of times the product is taken but not added to the shopping cart and the number of times the product is viewed, the product The number of inquiries, the number of times the product is taken, and the number of times the product is added to the shopping cart are spliced, and during the splicing process, the system does not save or record the original data of the customer, and does not pay attention to individual data, so it does not infringe User Privacy.

Fig. 2 is a second schematic flow chart of an intelligent design method for a personalized discount coupon according to an embodiment of the present application. Referring to Fig. 2, in some embodiments, after obtaining the video stream captured by the camera, the method further includes:

Step S201, identify the identity of the pedestrian through the image classification model, and determine whether the identity of the pedestrian is a service personnel or a customer; wherein, identifying the identity of the pedestrian can be realized by the uniform dress of the service personnel (for example, brightly colored tooling, uniform LOGO, etc.) In addition, among them, image classification model can adopt MoibleNets series model to realize, and wherein, MoibleNets series model includes MoibleNetV1, MoibleNetV2, MobileNetV3...etc., the MoibleNets series model that adopts in the present embodiment is lightweight model, not only parameter The amount is small and the amount of calculation is also small, so its running speed is fast, which is conducive to improving the training speed of the model;

Step S202, when the identity of a pedestrian is identified as a customer, generate a temporary ID of the customer, perform a pedestrian attribute recognition task (Pedestrian Attribute Recognition, PAR) on the customer, and add the extracted customer's age range feature and gender feature to the product sales Situation feature, when the customer leaves the shooting area, the temporary ID is deleted. Since the above steps do not save the user's original data and do not pay attention to individual data, it does not violate the user's privacy. It is easy to understand that the purpose of the human attribute recognition task is to mine the attribute information of pedestrians from the input image (for example, age interval features and gender features, etc.), and the high-level semantic information of pedestrians is obtained by recognition and mining; the available training methods include: Based on the RAP algorithm of deep learning, by using manually designed low-level features, such as HOG (Histogram of Oriented Gradient, also known as directional gradient histogram feature), SIFT (Scale-invariant feature transform, also known as scale-invariant feature transformation) , combined with the classification algorithm SVM and the conditional random field (CRF), it is convenient to train the human attribute recognition task, wherein, the prior art of the RAP algorithm, HOG, SIFT, classification algorithm SVM and CRF is known to those skilled in the art , so I won’t go into details one by one.

Fig. 3 is a schematic flow chart of the third intelligent design method of a personalized discount coupon according to the embodiment of the present application. Referring to Fig. 3, in the process of practical application, since pedestrians can move freely within the viewing angle of the camera, their moving speed and posture changes will change at any time, so it is impossible to guarantee that each frame of the video is clear and usable. In order to improve the recognition effect of features, in some embodiments, the identity of pedestrians is identified through the image classification model, and the identity of pedestrians is judged as service personnel or customers. , the method also includes:

Step S301, using the target detection model to detect the rectangular area of the pedestrian position appearing in the video stream picture, and output the target detection result; wherein, the target detection model is an existing technology in the field, and will not be repeated here;

Step S302, cutting the pedestrian image according to the target detection result, and performing quality evaluation on the cut pedestrian image through the quality evaluation module; wherein, the quality evaluation module is an existing technology in the field, and will not be repeated here;

Step S303, if the image of the pedestrian does not conform to the preset quality evaluation standard of the quality evaluation module, the identity information identification of the pedestrian is not performed.

FIG. 4 is a schematic flow diagram of obtaining the number of times a product is viewed according to Embodiment 1 of the present application. With reference to FIG. 4 , in some embodiments, obtaining the number of times a product is viewed includes the following steps:

Step S401, when it is detected that the temporary ID of the customer is the same, calculate the stay time of the customer in the target product area;

Step S402, if the stay time is greater than the preset stay time, and the customer's activity range is smaller than the preset space range, then the pedestrian stays in the previous time span, and the customer's activity range is within the preset time span. The circumscribed rectangle of the activity track within the dwell time is recorded as the number of visits; among them, the preset dwell time is set according to customer needs, and there is no specific limitation here;

Step S403, if the customer's activity range is larger than the preset spatial range, then the customer's activity range is the circumscribed rectangle of the customer's activity track within the stay time, which is recorded as the number of visits;

Fig. 5 is a schematic flow diagram of obtaining the number of times or the time of consultation of a product according to an embodiment of the present application. With reference to Fig. 5, in one embodiment, obtaining the number of times of consultation or the time of consultation of a product includes the following steps:

Step S501, calculate the distance between the customer and the service personnel, when the detected distance is less than the preset distance, time the initial communication time T1, and when the detected distance is equal to the preset distance, time the end communication time T2; The distance is set according to the needs of the user, and is not specifically limited here; in this embodiment, it is possible to distinguish whether it is a customer or a service person through step S201, and details will not be repeated here.

Step S502, calculate the time when the target product was consulted in the previous time span as the end communication time T2 minus the start communication time T1. For example, if the timing start communication time is T1: 10:00:00 in 2021, and the timing end communication time T2 is 2021 10:30:00 seconds, then the consultation time is 30 minutes.

Fig. 6 is a schematic flow chart of obtaining the number of times that the product is taken and the number of times that the product is added to the shopping cart according to an embodiment of the present application. Referring to Fig. Including the following steps:

Step S601, detecting the location of the shopping cart and the shelf. The shelf location is relatively fixed and can be marked manually. In addition, the number of shelves and the fine-grained labeling can be adjusted according to the scene.

Step S602, when it is detected that the distance between the pedestrian and the shelf position is less than the preset fixed threshold, start detecting the position of the human hand and track the position of the human hand in real time; where the preset fixed threshold is set according to the user's needs, and is not specifically limited here ;

Step S603, if it is detected that human hands enter the shelf area, record the number of times the item is picked up once;

Step S604, if it is detected that a human hand enters the shopping cart area, record the number of times items are put into the shopping cart once.

In order to solve the cold start problem of new commodities and improve the practicality, in one embodiment, after calculating the characteristics of commodity sales status according to the tracking results, the method further includes:

For a new product that has no characteristics of product sales status, the most similar product with product sales status characteristics is found through the inherent attributes of the new product. Among them, the inherent attributes of new commodities are category, price and brand. For example, the new product is "Jiaduobao", but Jiaduobao does not have the characteristics of product sales status. Through the inherent attributes of "Jiaduobao" (type, price, brand, etc.) to find the most similar product with the characteristics of product sales status, That is, "Wanglaoji"; from step S104 to step S106, a discount coupon for a new commodity is generated. Since the product sales status features in this embodiment are calculated according to the tracking results of the product in the previous time span, in some other embodiments, in order to improve the model effect, it is also possible to use multiple time spans. Sales status characteristics, for example, the time of the last three product sales.

Fig. 7 is a schematic diagram of the fourth flow chart of an intelligent design method for personalized discount coupons according to an embodiment of the present application. Referring to Fig. 7, in order to meet the needs of destocking or increasing sales in combination with business, in an embodiment, according to After the tracking result calculates the characteristics of the sales status of the commodity, the method further includes:

Step S701, calculating the product attention score according to the characteristics of the product sales status; wherein, the calculation formula of the product attention score is as follows:

Among them, Score _Item represents the product attention score, Item represents discounted products,

Represents the weight for different features, which can be manually set according to the actual situation, x _i represents the i-th product sales status feature, and n is the total number of product sales status features;

In step S702, products are screened according to the attention score and sales volume of the products. Among them, the formula for filtering products is as follows:

Among them, Sale _Item represents the sales volume of the product (Item), and Score _Item represents the attention score of the product.

In addition, considering that the purchase situation and the discount situation in the transaction will have an impact on the sales of the commodity, in some other embodiments, those skilled in the art can also analyze the user's purchase situation and the discount situation in the transaction through some software programs or algorithms. , realize the quantitative calculation of the user's discount sensitivity to obtain the discount and sales curve. Different commodity discounts and sales curves are different; Fig. 8 is a schematic diagram of commodity discounts and sales curves according to Embodiment 1 of the present application. With reference to Fig. 8, it can be seen that the sales volume increases rapidly with the increase of the discount at the initial stage, but after the discount is reduced to a certain degree Change tends to stagnate, and sales cease to increase due to insufficient supply or saturation of purchasing power. In actual scenarios, the discount cannot reach 100%. Selecting items that are sensitive to changes in discounts can lead to more dramatic results. Note that the discount and sales curve at this time only shows a general trend, so it cannot be used to accurately predict sales, and is only used to analyze the sensitivity of products to discounts.

In some embodiments, the regression model is a linear regression model, a random forest model, or a gradient boosting decision tree model. Of course, in some other embodiments, the regression model can also choose other models with better regression effects according to user needs, which will not be done here Specific limits.

Fig. 9 is a schematic flow chart of generating a discount coupon under the condition of profit maximization in an embodiment of the present application. Referring to Fig. 9, in some embodiments, generating a discount coupon under the condition of profit maximization includes:

Step S901, calculating the discount value when the profit is the largest; those skilled in the art can calculate the discount value when the profit is the largest through the formula of total sales and the formula of total profit; wherein, the formula of total sales is: total sales = sales volume *Price = sales volume * (original price - discount amount), and the formula for total profit is: total profit = sales volume * single piece profit = sales volume * (original price - discount amount - cost), because the model can predict the different discounts of products According to the above formula of total sales or total profit, the discount value corresponding to the maximum calculated value can be calculated;

Step S902, call a preset discount coupon template, the discount coupon template includes the page and appearance of the discount coupon; wherein, the page and appearance can be set according to user needs, and are not specifically limited here;

Step S903, write the attributes of the discount coupon, the discount date and the discount value calculated in step S901 into the corresponding position of the discount coupon template, generate a discount coupon, and issue the discount coupon to the customer through the membership system. Among them, the discount coupon attribute, discount date and discount value can be set according to user needs, and no specific limitations are made here;

Among them, the membership system is an existing technology, and some basic information of customers is recorded in the membership system, and the use of membership system data means that the use of such information has been authorized by the user; as shown in the following table:

名称name	来源source
客户唯一标识customer unique identifier	商品会员系统Commodity membership system
性别gender	商品会员系统Commodity membership system
年龄age	商品会员系统Commodity membership system
用户购买商品情况User purchases	历史交易数据Historical transaction data
用户交易情况user transactions	历史交易数据Historical transaction data

Among them, the unique identifier refers to the ID used to identify the user's identity in the membership system, the gender refers to the gender of the customer, the age refers to the age of the customer, and the user's purchase status refers to the identification, type and quantity of the product actually purchased by the user. etc.; user transaction status refers to the user's payment method, payment amount, preferential method and discount amount, etc.

In order to maximize the conversion rate of discount coupons, of course, in some other embodiments, combined with expert experience and industry knowledge, those skilled in the art can also use some software algorithms, programs or models to analyze the usage of coupons to obtain the usage rate of coupons. The relationship between the usage time, the validity period of the discount coupon, and the type of product, the setting method of the coupon validity period is written as a rule (that is, if A product uses C discount in B situation, then the validity period is D) to realize the selection of an appropriate discount for a specific product Coupon time attributes, increase the conversion rate of discount coupons, and increase product profits.

It should be noted that each of the above-mentioned modules may be a function module or a program module, and may be realized by software or by hardware. For the modules implemented by hardware, the above modules may be located in the same processor; or the above modules may be located in different processors in any combination.

This embodiment also provides an electronic device, including a memory and a processor, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

Optionally, the above-mentioned electronic device may further include a transmission device and an input-output device, wherein the transmission device is connected to the above-mentioned processor, and the input-output device is connected to the above-mentioned processor.

Optionally, in this embodiment, the above-mentioned processor may be configured to execute the following steps through a computer program:

Step S101, obtaining the video stream captured by the camera;

Step S102, performing target tracking tasks on pedestrians and commodities in the video stream to obtain tracking results;

Step S103, calculating the product sales status features according to the tracking results, the product sales status features at least including the number of times the product is viewed, the number of times the product is consulted, the number of times the product is taken, and the number of times the product is added to the shopping cart;

Step S104, using the characteristics of the commodity sales status to train a regression model to obtain a sales forecast model;

Step S105, using the sales forecast model to predict the sales volume of the commodity under a given discount;

Step S106, according to the predicted sales volume, a discount coupon is generated under the condition of maximizing profit.

In addition, in combination with the intelligent design method for personalized discount coupons in the foregoing embodiments, the embodiments of the present application may provide a storage medium for implementation. A computer program is stored on the storage medium; when the computer program is executed by a processor, any intelligent design method for a personalized discount coupon in the above-mentioned embodiments is realized.

In one embodiment, a computer device is provided, and the computer device may be a terminal. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Wherein, the processor of the computer device is used to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by a processor, an intelligent design method for personalized discount coupons is realized. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a button, a trackball or a touch pad provided on the casing of the computer device , and can also be an external keyboard, touchpad, or mouse.

In one embodiment, FIG. 10 is a schematic diagram of the internal structure of an electronic device according to an embodiment of the present application. As shown in FIG. 10 , an electronic device is provided. shown. The electronic device includes a processor connected through an internal bus, a network interface, an internal memory and a non-volatile memory, wherein the non-volatile memory stores an operating system, a computer program and a database. The processor is used to provide computing and control capabilities, the network interface is used to communicate with external terminals through a network connection, and the internal memory is used to provide an environment for the operation of the operating system and computer programs. When the computer program is executed by the processor, a personalized Smart design approach for discount coupons, database is used to store data.

Those skilled in the art can understand that the structure shown in Figure 10 is only a block diagram of a part of the structure related to the solution of this application, and does not constitute a limitation on the electronic device on which the solution of this application is applied. The specific electronic device can be More or fewer components than shown in the figures may be included, or some components may be combined, or have a different arrangement of components.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be realized by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium , when the computer program is executed, it may include the procedures of the embodiments of the above-mentioned methods. Wherein, any references to memory, storage, database or other media used in the various embodiments provided in the present application may include non-volatile and/or volatile memory. Nonvolatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Chain Synchlink DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Those skilled in the art should understand that the various technical features of the above-mentioned embodiments can be combined arbitrarily. There is no contradiction in the combination of technical features, and all should be considered as within the scope of the description.

The above-mentioned embodiments only represent several implementation modes of the present application, and the description thereof is relatively specific and detailed, but it should not be construed as limiting the scope of the patent for the invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the scope of protection of the patent application should be based on the appended claims.

Claims

An intelligent design method for personalized discount coupons, characterized in that the method comprises the following steps:

Obtain the video stream captured by the camera;

Perform target tracking tasks on pedestrians and commodities in the video stream to obtain tracking results;

According to the tracking results, the product sales status feature is calculated, and the product sales status feature at least includes the number of times the product is viewed, the number of times the product is consulted, the number of times the product is taken, and the number of times the product is added to the shopping cart;

Utilize the characteristic training regression model of described commodity sales situation, obtain sales volume prediction model;

Using the sales forecast model to predict the sales volume of the commodity under a given discount;

According to the predicted sales volume, discount coupons are generated under the condition of maximizing profit.
The method according to claim 1, wherein the product sales status feature further includes the number of times the product has been browsed for more than a preset threshold but has not been taken, and the number of times the product has been taken but not added to the shopping cart.
The method according to claim 1, characterized in that, after the acquisition of the video stream captured by the camera, the method further comprises:

Identify the identity of the pedestrian through the image classification model, and judge whether the identity of the pedestrian is a service person or a customer;

When the identity of a pedestrian is identified as a customer, generate the temporary ID of the customer, and perform the pedestrian attribute identification task on the customer, and add the extracted customer's age interval feature and gender feature to the commodity sales status feature, when the When the customer leaves the shooting area, the temporary ID is deleted.
The method according to claim 3, wherein, before identifying the identity of the pedestrian through the image classification model and judging whether the identity of the pedestrian is a service personnel or a customer, the method further comprises:

Detecting the rectangular area of the pedestrian position appearing in the video stream picture through the target detection model, and outputting the target detection result;

cutting the pedestrian image according to the target detection result, and performing quality evaluation on the cut pedestrian image through the quality evaluation module;

If the image of the pedestrian does not conform to the preset quality evaluation standard of the quality evaluation module, the identity information of the pedestrian is not identified.
The method according to claim 1, characterized in that, after calculating the commodity sales status characteristics according to the tracking results, the method further comprises:

For a new product without the characteristics of the sales status of the product, the most similar product with the characteristics of the sales status of the product is found through the inherent attributes of the new product.
The method according to claim 1, characterized in that, after calculating the commodity sales status characteristics according to the tracking results, the method further comprises:

Calculating the commodity attention score according to the characteristics of the commodity sales status;

Products are screened according to the attention score and sales volume of the products.
The method according to any one of claims 1-6, wherein the regression model is a linear regression model, a random forest model or a gradient boosting decision tree model.
The method according to claim 1, wherein said generating discount coupons under profit maximization conditions comprises:

Calculation of the discount value at which the profit is maximized;

Call the preset discount coupon template, which includes the page and appearance of the discount coupon;

Write the discount coupon attribute, discount date and the discount value into the corresponding position of the discount coupon template, generate the discount coupon, and issue the discount coupon to the customer through the membership system.
An electronic device, comprising a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to perform the method described in any one of claims 1 to 8 A smart design approach to personalizing discount coupons.
A storage medium, characterized in that a computer program is stored in the storage medium, wherein the computer program is configured to execute the intelligent design of the personalized discount coupon described in any one of claims 1 to 8 when running method.