Facial big data efficient management system for physical beauty store

Abstract

The invention provides a face big data efficient management system for a physical beauty store, and relates to the technical field of digital beauty. According to the invention, the terminal information acquisition unit automatically acquires the beauty related information retained in the whole process of the user in the store, and the store information acquisition unit automatically acquires the ERP purchase-sell-stock information, so that the data barrier is broken through and the problem of key information omission is avoided; the user is subjected to tagging analysis through the data tag analysis unit, so that accurate evaluation of the user demand is realized; the user archive management unit can integrate and utilize user data and ensure the security and integrity of information; the demand analysis pushing module can carry out targeted and accurate pushing according to the user demand and the self condition of the store, and a shop assistant does not need to carry out user demand analysis; according to the whole scheme, dependence on personnel is greatly reduced, the employee service threshold of the physical store is reduced, store management can be copied, and the physical store is helped to develop and expand.

Description

Technical Field

[0001] This invention relates to the field of digital beauty technology, and in particular to a highly efficient facial big data management system for physical beauty salons. Background Technology

[0002] With the advancement of technology and the development of the times, people's pursuit of beauty is no longer limited to simple skin protection. To this end, invention patent application number CN202110886352.1 proposes a portable skin quality detection device and system based on image recognition. This application allows for convenient and quick skin quality detection via tablet or mobile phone. The detection report can be compared horizontally through historical records to understand the skin's improvement, helping users track their skin condition and choose suitable skincare products. By analyzing skin quality, skincare products can be customized for different areas. The customized raw materials are provided to consumers through a skincare product customization machine, and consumers then apply the skincare products to the corresponding areas, thus achieving the goal of providing customized skincare products according to the needs of different skin areas.

[0003] The aforementioned application has been implemented in physical beauty salons, where staff use mobile phones and tablets to access a WeChat mini-program, collect customers' facial images, generate personalized facial analysis reports, and provide customized skincare products based on the analysis reports to meet the different care needs of various skin areas. However, this application only allows for one-time services to customers and does not mine or track the relevant data, resulting in the data value not being fully utilized and the inability to accurately understand users' care needs.

[0004] Therefore, in view of the shortcomings of existing technologies in data management, this application provides an efficient facial big data management system for physical beauty stores to solve the above-mentioned technical problems. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a highly efficient facial big data management system for physical beauty salons, comprising: a terminal information acquisition unit, a store information acquisition unit, a data tag analysis unit, a user profile management unit, and a demand analysis and push module; wherein,

[0006] The terminal information acquisition unit is deployed on various terminal devices in physical beauty stores. With user authorization, it acquires user-related beauty information through these terminal devices. This includes: a user access control acquisition module, a user voice acquisition module, a user reading acquisition module, a user evaluation acquisition module, and a user report acquisition module. The store information acquisition unit categorizes and organizes the physical beauty store's ERP inventory information to obtain cosmetic purchase information, cosmetic sales information, and cosmetic inventory information.

[0007] The data tagging analysis unit performs tagging analysis on users' beauty-related information and cosmetics sales information to obtain primary tags; then, it constructs user profiles based on primary tags of different dimensions and tags them to obtain deep tags; the user profile management unit establishes personal data profiles for users and tracks and stores users' beauty-related information, primary tags, and deep tags.

[0008] The demand analysis and push module analyzes and predicts users' beauty needs at different times based on beauty-related information, primary tags, and deep tags in their personal data profiles, obtains the predicted demand for beauty products, and pushes products based on the predicted demand.

[0009] As a further solution, the primary tags are divided into communication interest tags, reading interest tags, and evaluation interest tags on the interest side, and the primary tags on the interest side are fused to obtain interest-related deep tags; the primary tags are divided into consumption cycle tags, consumption level tags, consumption category tags, and consumption trajectory tags on the consumption side, and the primary tags on the consumption side are fused to obtain consumption habit deep tags; user profiles are constructed by using interest-related deep tags and consumption habit deep tags, and the predicted demand of users in various scenarios is estimated based on user profiles.

[0010] As a further solution, after obtaining information collection permissions through the user access control module, the terminal information acquisition unit acquires the user's beauty-related information throughout the entire process of the user receiving services at a physical beauty salon through the following steps:

[0011] When a user undergoes a skin analysis, the user's voice data is acquired through a user voice acquisition module. This module collects data from the microphones of various terminal devices in the physical beauty store and records the data when beauty-related keywords are detected, thus obtaining the user's voice data.

[0012] When users read the test report, user reading data is obtained through the user reading acquisition module; wherein, the user reading acquisition module obtains user reading data by performing reading tests on various terminal devices in physical beauty stores and obtaining reading report content that users are interested in through content detection.

[0013] After receiving services from a store, users can obtain user feedback through a questionnaire via a user feedback acquisition module. The user feedback acquisition module extracts content from the user feedback and quantifies and categorizes the user feedback based on the questionnaire content to obtain user feedback data.

[0014] As a further solution, the user reading acquisition module includes a reading window module, an eye-tracking module, and a reading statistics module, which are deployed in the background on various terminal devices in physical beauty stores. When a user reads the test report through the terminal device, the reading window module is called in the background to obtain the current page of the user's reading report, the eye-tracking module obtains the content the user is gazing at on the current page, and the reading statistics module obtains the dwell time and number of times the user is gazing at the content. By performing tag analysis on the dwell time, number of times, and gazing content, reading attention tags are obtained.

[0015] As a further solution, the user voice data is processed through voice transcoding, data cleaning, and preprocessing to extract voice-text data; the voice-text data is then converted into voice-text feature vectors and input into a beauty tag extraction model to obtain the user's communication and attention tags; the user review data is processed through data cleaning and preprocessing to extract review text data; the review text data is then converted into review text feature vectors and input into a beauty tag extraction model to obtain the user's review and attention tags.

[0016] The data cleaning includes removing irrelevant, duplicate, and erroneous information; the preprocessing includes removing punctuation marks, word equivalence conversion, and text standardization; the beauty label extraction model uses text data already labeled with beauty labels as a training set and performs positive / negative training to obtain an extraction model that meets the recognition accuracy requirements.

[0017] As a further solution, cosmetic sales information is obtained through an ERP inventory management system, and user consumption data is summarized. This user consumption data includes historical data on user purchase time, purchase amount, purchased products, and purchased stores. By analyzing and summarizing historical purchase time data, user consumption cycle tags are obtained; by statistically summarizing historical purchase amount data, user consumption level tags are obtained; by classifying and summarizing purchased products, user consumption category tags are obtained; and by summarizing the location of purchased stores, user consumption trajectory tags are obtained.

[0018] As a further solution, demand-driven push includes pushing beauty products to users; this is done through the following steps:

[0019] Step A1: The store information acquisition unit connects to the store's EPR inventory management system and obtains the cosmetics inventory information of the physical beauty store to get the current inventory of cosmetics products in the store.

[0020] Step A2: Obtain information on stocked beauty products, including sales time, product price, product type, and store location;

[0021] Step A3: Input the inventory of beauty products into the user profile and estimate the predicted demand of users for the inventory of beauty products.

[0022] Step A4: Set users whose predicted demand meets the potential purchase threshold as potential users and push the current inventory of beauty products to potential users.

[0023] As a further solution, demand push includes sending beauty product requests to stores; this is done through the following steps:

[0024] Step B1: The store information acquisition unit connects to the store's EPR inventory management system and obtains the cosmetics purchase information of the physical beauty store to get the cosmetics products currently on sale in the store.

[0025] Step B2: Obtain information on available beauty products, including sales hours, product prices, product types, and store locations.

[0026] Step B3: Input the information of the beauty products on sale into the user profile and estimate the predicted demand of users for the beauty products on sale;

[0027] Step B4: Set users whose predicted demand meets the potential purchase threshold as potential users, and count the number of potential users;

[0028] Step B5: Push product sales to potential users, obtain pre-purchase demand, and adjust the beauty product procurement information of the products on sale according to beauty needs.

[0029] As a further solution, the user profile management unit also parses user reports to obtain various indicators of user skin health, and generates corresponding indicator tracking reports based on these indicators.

[0030] As a further solution, the data tag analysis unit, user profile management unit, and demand analysis push module are all deployed in the cloud and centrally managed through the cloud platform.

[0031] Compared with related technologies, the efficient facial big data management system for physical beauty salons provided by this invention has the following beneficial effects:

[0032] This invention automatically acquires beauty-related information stored by users throughout their time in the store through a terminal information acquisition unit, and automatically retrieves ERP inventory information through a store information acquisition unit, breaking down data barriers and avoiding the omission of key information. A data tagging analysis unit performs tag-based analysis on users, enabling accurate assessment of user needs. A user profile management unit integrates and utilizes user data while ensuring information security and integrity. A demand analysis and push module provides targeted and precise pushes based on user needs and the store's own situation, eliminating the need for store staff to perform user demand analysis. The entire solution significantly reduces reliance on personnel and lowers the service threshold for physical store employees, making store management replicable and thus helping physical stores grow and expand. Attached Figure Description

[0033] Figure 1 A schematic diagram of the structure of a high-efficiency facial big data management system for physical beauty salons provided by the present invention;

[0034] Figure 2 This is a schematic diagram of a user skin detection image provided in an embodiment of the present invention;

[0035] Figure 3 This is a schematic diagram of a user skin detection report provided in an embodiment of the present invention;

[0036] Figure 4 This is a schematic diagram of a user profile management unit provided in an embodiment of the present invention;

[0037] Figure 5 This is a schematic diagram of a data tag analysis unit provided in an embodiment of the present invention;

[0038] Figure 6 This is a schematic diagram of a store label management interface provided in an embodiment of the present invention;

[0039] Figure 7 This is a schematic diagram of store beauty product information provided in an embodiment of the present invention;

[0040] Figure 8 This is a schematic diagram illustrating the push notification of beauty products to users, provided in an embodiment of the present invention.

[0041] Figure 9 A schematic diagram illustrating personal data profile management is provided for embodiments of the present invention;

[0042] Figure 10 This is a schematic diagram of some indicator tracking reports provided in an embodiment of the present invention;

[0043] Figure 11 This invention provides a schematic diagram of potential users pre-purchasing products. Detailed Implementation

[0044] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0045] like Figure 1 As shown in the figure, this embodiment provides a high-efficiency facial big data management system for physical beauty salons, including: a terminal information acquisition unit, a store information acquisition unit, a data tag analysis unit, a user profile management unit, and a demand analysis and push module; wherein,

[0046] The terminal information acquisition unit is deployed on various terminal devices in physical beauty stores. With user authorization, it acquires user-related beauty information through these terminal devices. This includes: a user access control acquisition module, a user voice acquisition module, a user reading acquisition module, a user evaluation acquisition module, and a user report acquisition module. The store information acquisition unit categorizes and organizes the physical beauty store's ERP inventory information to obtain cosmetic purchase information, cosmetic sales information, and cosmetic inventory information.

[0047] The data tagging analysis unit performs tagging analysis on users' beauty-related information and cosmetics sales information to obtain primary tags; then, it constructs user profiles based on primary tags of different dimensions and tags them to obtain deep tags; the user profile management unit establishes personal data profiles for users and tracks and stores users' beauty-related information, primary tags, and deep tags.

[0048] The demand analysis and push module analyzes and predicts users' beauty needs at different times based on beauty-related information, primary tags, and deep tags in their personal data profiles, obtains the predicted demand for beauty products, and pushes products based on the predicted demand.

[0049] It should be noted that existing technologies (such as the invention patent with application number CN202110886352.1) have proposed methods for convenient and quick skin quality testing via tablets or mobile phones; however, in actual use, we have found that: 1. It is difficult to cultivate the professionalism of employees, and often the needs of users are not fully explored and developed, making it difficult for store staff to recommend products based on customer needs; 2. User information management is disorganized and cannot be integrated and utilized, sometimes resulting in the omission of key information and data security issues; 3. Store marketing management is inaccurate, making it impossible to accurately assess the relationship between user needs and the store's own products; 4. Store management is difficult to replicate, thus restricting development and expansion.

[0050] To address this, this embodiment utilizes a highly efficient management system based on terminal devices and an ERP inventory management system. The terminal information acquisition unit automatically retrieves beauty-related information stored throughout a user's visit to the store, while the store information acquisition unit automatically acquires ERP inventory information, breaking down data barriers and preventing the omission of crucial information. The data tagging analysis unit performs tag-based analysis on users, enabling accurate assessment of user needs. The user profile management unit integrates and utilizes user data while ensuring information security and integrity. The demand analysis and push module provides targeted and precise pushes based on user needs and the store's specific circumstances, eliminating the need for staff to perform user demand analysis and achieving a "from novice to expert, from expert to professor" effect. This entire solution significantly reduces reliance on personnel and lowers the service threshold for physical stores, making store management replicable and thus helping physical stores grow and expand.

[0051] The following diagrams illustrate this scheme in detail:

[0052] In physical stores, users input information via terminal devices, such as... Figure 2 The face detection image shown is used to generate a face detection image, and the generated face detection image is generated as shown in the image. Figure 3 The skin analysis report shown allows users to view their skin condition and any existing skin problems. With the user's authorization, the store collects beauty-related information from the user and uses methods such as... Figure 4 The user profile management unit shown manages user data, and the data tag analysis unit can perform tag-based analysis based on the user data to obtain results such as... Figure 5 The various primary tags shown are used to construct user profiles and tagged them with different aspects of primary tags, resulting in deep tags that accurately depict user characteristics; stores can then use such... Figure 6 The store tag management interface shown filters users who meet the corresponding tag characteristics.

[0053] Based on this, stores obtain cosmetics sales information through the ERP inventory management system, and summarize the data to obtain user consumption data, through methods such as... Figure 7 The product information management interface shown allows users to view and edit relevant data for consumer products. After analyzing user needs through the demand analysis and push module, the system can identify the user's target product or the product's target customer group, and then... Figure 8 The user beauty product push interface shown allows for precise product recommendations to users, thus completing a targeted and accurate push based on user needs and the store's own situation.

[0054] As a further solution, the primary tags are divided into communication interest tags, reading interest tags, and evaluation interest tags on the interest side, and the primary tags on the interest side are fused to obtain interest-related deep tags; the primary tags are divided into consumption cycle tags, consumption level tags, consumption category tags, and consumption trajectory tags on the consumption side, and the primary tags on the consumption side are fused to obtain consumption habit deep tags; user profiles are constructed by using interest-related deep tags and consumption habit deep tags, and the predicted demand of users in various scenarios is estimated based on user profiles.

[0055] It should be noted that this embodiment uses user tagging to accurately model user profiles, thereby achieving the effect of accurately predicting user needs. The first-level tags on the interest side reflect the beauty objects that users are interested in, and we can target these objects (eye bags, blackheads, wrinkles, etc.) with corresponding products to meet users' beauty needs. The first-level tags on the consumption side reflect users' consumption habits, and we can conduct targeted marketing based on these habits to meet users' consumption needs (meeting both product needs and consumption needs).

[0056] As a further solution, after obtaining information collection permissions through the user access control module, the terminal information acquisition unit acquires the user's beauty-related information throughout the entire process of the user receiving services at a physical beauty salon through the following steps:

[0057] When a user undergoes a skin analysis, the user's voice data is acquired through a user voice acquisition module. This module collects data from the microphones of various terminal devices in the physical beauty store and records the data when beauty-related keywords are detected, thus obtaining the user's voice data.

[0058] When users read the test report, user reading data is obtained through the user reading acquisition module; wherein, the user reading acquisition module obtains user reading data by performing reading tests on various terminal devices in physical beauty stores and obtaining reading report content that users are interested in through content detection.

[0059] After receiving services from a store, users can obtain user feedback through a questionnaire via a user feedback acquisition module. The user feedback acquisition module extracts content from the user feedback and quantifies and categorizes the user feedback based on the questionnaire content to obtain user feedback data.

[0060] It should be noted that this embodiment uses a terminal device for one-stop data collection, ensuring a seamless user experience without affecting service quality, while collecting comprehensive information. In particular, the ability to perceive user interests through reading the test report is highly innovative. Other methods of obtaining interests essentially involve mining relevant text, extracting key information, and then using a beauty tag extraction model to obtain the user's reading / evaluation and attention tags.

[0061] As a further solution, the user reading acquisition module includes a reading window module, an eye-tracking module, and a reading statistics module, which are deployed in the background on various terminal devices in physical beauty stores. When a user reads the test report through the terminal device, the reading window module is called in the background to obtain the current page of the user's reading report, the eye-tracking module obtains the content the user is gazing at on the current page, and the reading statistics module obtains the dwell time and number of times the user is gazing at the content. By performing tag analysis on the dwell time, number of times, and gazing content, reading attention tags are obtained.

[0062] As a further solution, the user voice data is processed through voice transcoding, data cleaning, and preprocessing to extract voice-text data; the voice-text data is then converted into voice-text feature vectors and input into a beauty tag extraction model to obtain the user's communication and attention tags; the user review data is processed through data cleaning and preprocessing to extract review text data; the review text data is then converted into review text feature vectors and input into a beauty tag extraction model to obtain the user's review and attention tags.

[0063] The data cleaning includes removing irrelevant, duplicate, and erroneous information; the preprocessing includes removing punctuation marks, word equivalence conversion, and text standardization; the beauty label extraction model uses text data already labeled with beauty labels as a training set and performs positive / negative training to obtain an extraction model that meets the recognition accuracy requirements.

[0064] As a further solution, cosmetic sales information is obtained through an ERP inventory management system, and user consumption data is summarized. This user consumption data includes historical data on user purchase time, purchase amount, purchased products, and purchased stores. By analyzing and summarizing historical purchase time data, user consumption cycle tags are obtained; by statistically summarizing historical purchase amount data, user consumption level tags are obtained; by classifying and summarizing purchased products, user consumption category tags are obtained; and by summarizing the location of purchased stores, user consumption trajectory tags are obtained.

[0065] It should be noted that this embodiment integrates the store's beauty big data and EPR inventory management system. By obtaining beauty sales information through the EPR inventory management system and summarizing the data, user consumption data can be obtained. This can accurately and objectively reflect users' consumption habits, consumption needs, consumption events, and price ranges, thereby better meeting users' consumption needs.

[0066] As a further solution, demand-driven push includes pushing beauty products to users; this is done through the following steps:

[0067] Step A1: The store information acquisition unit connects to the store's EPR inventory management system and obtains the cosmetics inventory information of the physical beauty store to get the current inventory of cosmetics products in the store.

[0068] Step A2: Obtain information on stocked beauty products, including sales time, product price, product type, and store location;

[0069] Step A3: Input the inventory of beauty products into the user profile and estimate the predicted demand of users for the inventory of beauty products.

[0070] Step A4: Set users whose predicted demand meets the potential purchase threshold as potential users and push the current inventory of beauty products to potential users.

[0071] It should be noted that our demand push can perform two-way demand matching. After constructing user profiles, this embodiment can push beauty products to users based on their product needs; for example... Figure 11 As shown, we can provide users with the products they need in a timely manner, thereby improving service quality and efficiency and increasing user stickiness.

[0072] As a further solution, demand push includes sending beauty product requests to stores; this is done through the following steps:

[0073] Step B1: The store information acquisition unit connects to the store's EPR inventory management system and obtains the cosmetics purchase information of the physical beauty store to get the cosmetics products currently on sale in the store.

[0074] Step B2: Obtain information on available beauty products, including sales hours, product prices, product types, and store locations.

[0075] Step B3: Input the information of the beauty products on sale into the user profile and estimate the predicted demand of users for the beauty products on sale;

[0076] Step B4: Set users whose predicted demand meets the potential purchase threshold as potential users, and count the number of potential users;

[0077] Step B5: Push product sales to potential users, obtain pre-purchase demand, and adjust the beauty product procurement information of the products on sale according to beauty needs.

[0078] It should be noted that our demand push can perform two-way demand matching. After obtaining information from the EPR inventory management system, this embodiment can push beauty demand to stores based on inventory status, and adjust beauty purchase information in a timely manner to ensure that inventory does not accumulate.

[0079] As a further solution, the user profile management unit also parses user reports to obtain various indicators of user skin health, and generates corresponding indicator tracking reports based on these indicators.

[0080] It should be noted that this embodiment can address situations such as... Figure 10 The system provides indicator tracking reports, which objectively reflect the user's skin improvement status, enabling them to better select more effective beauty products.

[0081] As a further solution, the data tag analysis unit, user profile management unit, and demand analysis push module are all deployed in the cloud and centrally managed through the cloud platform.

[0082] It should be noted that traditional brick-and-mortar store customer management is usually maintained and managed by employees themselves. While this allows employees to sell products based on customer needs, it also poses challenges to store data security. This embodiment addresses this by... Figure 9 The cloud platform shown centrally manages user profiles, user data, and user needs, thereby achieving "data flow when people leave," ensuring data security for physical stores and preventing customer churn.

[0083] In summary, we proposed a series of solutions to address the pain points faced by brick-and-mortar stores. Firstly, by partnering with platforms like Douyin and Meituan, we shifted our focus from relying on product price-based customer acquisition to utilizing online tools to attract a wider user base. Secondly, through a human-machine collaboration approach, we trained ordinary employees to become experts and even professors, thus solving the problem of training beauty advisors (BAs) with specialized skills and improving training efficiency. Furthermore, to address the issue of customer churn due to staff turnover, we avoid directly adding customers on WeChat. Instead, we retain data and ensure store security through a unified backend information management system.

[0084] To address the pain point of inaccurate membership marketing management, we employ a "tackling the threads" approach, tagging each customer to achieve intelligent, precise marketing. To address the issue of homogenized experiential products, we utilize AI algorithms to provide more accurate product matching and implement an offline delivery model to meet precise consumer needs. Furthermore, through membership data collection, report reading, surveys, and mini-program appointments, as well as integrating with our online store, we have achieved a transformation from having no consumer information tracking to using AI for big data-driven membership information management.

[0085] To address the security concerns of membership and sales data, we rely on Alibaba Cloud and data regulations for protection. Furthermore, by establishing a unique physical store operation model and standardizing product formulations for each store, we prevent standardized products from being vulnerable to online competition. Addressing the pain point of the inability to integrate facial skin data with the ERP system, we have implemented true data-driven analysis and reporting to proactively address potential issues. Finally, to overcome the challenge of replicating branch openings, we rely on and support the platform and modern management tools to enable easily replicable expansion.

[0086] In summary, by providing solutions to address each pain point, we can effectively improve business operation models, enhance marketing accuracy and efficiency, achieve data-driven management, protect the security of member and sales data, and promote business replicability and sustainable development.

[0087] The above are merely embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A facial big data efficient management system for a physical beauty store, characterized by, include: The system includes a terminal information acquisition unit, a store information acquisition unit, a data tag analysis unit, a user profile management unit, and a demand analysis and push module; among which, The terminal information acquisition unit is deployed on various terminal devices in physical beauty stores. With user authorization, it acquires user-related beauty information through these terminal devices. This includes: a user access control acquisition module, a user voice acquisition module, a user reading acquisition module, a user evaluation acquisition module, and a user report acquisition module. The store information acquisition unit categorizes and organizes the physical beauty store's ERP inventory information to obtain cosmetic purchase information, cosmetic sales information, and cosmetic inventory information. The data tagging analysis unit performs tagging analysis on users' beauty-related information and cosmetics sales information to obtain primary tags; then, it constructs user profiles based on primary tags of different dimensions and tags them to obtain deep tags; the user profile management unit establishes personal data profiles for users and tracks and stores users' beauty-related information, primary tags, and deep tags. The demand analysis and push module analyzes and predicts users' beauty needs at different times based on beauty-related information, primary tags, and deep tags in their personal data profiles, obtains the predicted demand for beauty products, and pushes products based on the predicted demand.

2. The efficient facial big data management system for physical beauty salons according to claim 1, characterized in that, The primary tags are divided into communication interest tags, reading interest tags, and evaluation interest tags on the interest side. The primary tags on the interest side are then fused to obtain interest-related deep tags. On the consumption side, the primary tags are divided into consumption cycle tags, consumption level tags, consumption category tags, and consumption trajectory tags. The primary tags on the consumption side are then fused to obtain consumption habit deep tags. User profiles are constructed by combining interest-related deep tags and consumption habit deep tags, and the predicted demand of users in various scenarios is estimated based on these user profiles.

3. The efficient facial big data management system for physical beauty salons according to claim 2, characterized in that, After obtaining information collection permissions through the user access control module, the terminal information acquisition unit acquires the user's beauty-related information throughout the entire process of the user receiving services at a physical beauty salon through the following steps: When a user undergoes a skin analysis, the user's voice data is acquired through a user voice acquisition module. This module collects data from the microphones of various terminal devices in the physical beauty store and records the data when beauty-related keywords are detected, thus obtaining the user's voice data. When users read the test report, user reading data is obtained through the user reading acquisition module; wherein, the user reading acquisition module obtains user reading data by performing reading tests on various terminal devices in physical beauty stores and obtaining reading report content that users are interested in through content detection. After receiving services from a store, users can obtain user feedback through a questionnaire via a user feedback acquisition module. The user feedback acquisition module extracts content from the user feedback and quantifies and categorizes the user feedback based on the questionnaire content to obtain user feedback data.

4. The efficient facial big data management system for physical beauty salons according to claim 3, characterized in that, The user reading acquisition module includes a reading window module, an eye-tracking module, and a reading statistics module, and is deployed in the background on various terminal devices in physical beauty stores. When a user reads a test report through a terminal device, the background calls the reading window module to obtain the current page of the user's reading report, the eye-tracking module to obtain the content the user is gazing at on the current page, and the reading statistics module to obtain the dwell time and number of times the user is gazing at the content. By performing tag analysis on the dwell time, number of times, and gazing content, reading attention tags are obtained.

5. The efficient facial big data management system for physical beauty salons according to claim 3, characterized in that, The user's voice data is extracted through voice transcoding, data cleaning, and preprocessing to obtain voice-text data; the voice-text data is then converted into voice-text feature vectors and input into the beauty tag extraction model to obtain the user's communication and interest tags; The user review data is cleaned and preprocessed to extract review text data; the review text data is then converted into review text feature vectors and input into the beauty tag extraction model to obtain the user's review attention tags. The data cleaning includes removing irrelevant, duplicate, and erroneous information; the preprocessing includes removing punctuation marks, word equivalence conversion, and text standardization; the beauty label extraction model uses text data with pre-labeled beauty labels as a training set and performs positive / negative training to obtain an extraction model that meets the recognition accuracy requirements.

6. The efficient facial big data management system for physical beauty salons according to claim 2, characterized in that, The system obtains cosmetics sales information through ERP inventory management and summarizes the data to obtain user consumption data. The user consumption data includes historical data on user purchase time, purchase amount, purchased products, and purchased stores. By analyzing and summarizing historical purchase time data, the system obtains user consumption cycle tags; by statistically summarizing historical purchase amount data, the system obtains user consumption level tags; by classifying and summarizing purchased products, the system obtains user consumption category tags; and by summarizing the location of purchased stores, the system obtains user consumption trajectory tags.

7. The efficient facial big data management system for physical beauty salons according to claim 2, characterized in that, Demand-driven push notifications include sending beauty product recommendations to users; these recommendations are implemented through the following steps: Step A1: The store information acquisition unit connects to the store's EPR inventory management system and obtains the cosmetics inventory information of the physical beauty store to get the current inventory of cosmetics products in the store. Step A2: Obtain information on stocked beauty products, including sales time, product price, product type, and store location; Step A3: Input the inventory of beauty products into the user profile and estimate the predicted demand of users for the inventory of beauty products. Step A4: Set users whose predicted demand meets the potential purchase threshold as potential users and push the current inventory of beauty products to potential users.

8. The efficient facial big data management system for physical beauty salons according to claim 2, characterized in that, Demand push includes sending beauty product requests to stores; the beauty product request push is carried out through the following steps: Step B1: The store information acquisition unit connects to the store's EPR inventory management system and obtains the cosmetics purchase information of the physical beauty store to get the cosmetics products currently on sale in the store. Step B2: Obtain information on available beauty products, including sales hours, product prices, product types, and store locations. Step B3: Input the information of the beauty products on sale into the user profile and estimate the predicted demand of users for the beauty products on sale; Step B4: Set users whose predicted demand meets the potential purchase threshold as potential users, and count the number of potential users; Step B5: Push product sales to potential users, obtain pre-purchase demand, and adjust the beauty product procurement information of the products on sale according to beauty needs.

9. The efficient facial big data management system for physical beauty salons according to claim 1, characterized in that, The user profile management unit also parses user reports to obtain various indicators of user skin health, and generates corresponding indicator tracking reports based on these indicators.

10. The efficient facial big data management system for physical beauty salons according to claim 1, characterized in that, The data tag analysis unit, user profile management unit, and demand analysis and push module are all deployed in the cloud and centrally managed through the cloud platform.

Images