Intelligent retail method and system based on face recognition

Abstract

The invention relates to the technical field of retail intelligent management. The system comprises a moving track analysis module, a region division module, a commodity association analysis module and a commodity recombination module. The movement track analysis module is used for analyzing and judging the abnormal behavior track data by tracking the movement track of the shopping basket and generating a behavior thermodynamic diagram according to the discrete movement track data; the region division module divides the retail space into a hot region, a warm region and a cold region; the commodity association analysis module analyzes the association values of the association combinations and constructs a commodity association graph according to the association combinations and the corresponding association values; and the commodity recombination module recombines the commodities in the hot area and the cold area in the retail space through the strong correlation commodity group. According to the method, the correlation degree of high-frequency interaction and low-frequency interaction commodities in the retail space is analyzed, and the commodities are scientifically recombined, so that the retail operation efficiency is remarkably improved, and the crowded condition is avoided.

Description

Technical Field

[0001] The present invention relates to the field of retail intelligent management technology, and in particular to an intelligent retail method and system based on face recognition. Background Art

[0002] Traditional retail stores face multiple challenges, including changing consumer behavior, the impact of e-commerce, and rising operating costs. To address these challenges, smart retail stores have emerged. By integrating advanced technologies such as the Internet of Things (IoT), big data, artificial intelligence (AI), and machine learning, they are transforming traditional offline merchandise transactions into smart retail, leveraging internet, IoT, and AI technologies. With the rapid advancement of technology, the retail industry is undergoing unprecedented transformation. These improvements are enhancing the shopping experience, optimizing operational efficiency, and strengthening data analysis capabilities, thereby meeting consumers' increasingly personalized needs and boosting retailers' competitiveness.

[0003] As a core component of smart retail stores, user trajectory tracking technology provides retailers with in-depth insights by collecting and analyzing real-time data such as consumers' movement paths, dwell time, and interactive behaviors in the store to support precision marketing, optimize product display, and improve customer satisfaction and operational efficiency.

[0004] Under existing technologies, user trajectory tracking technology can only show consumers' physical movements within the store, but cannot directly reflect their purchasing intentions or actual interest in products, nor can it reveal the correlation between consumers and different products. Relying solely on user trajectory tracking technology analysis will lead to errors in the behavioral heat map of retail stores, thereby reducing customer satisfaction and operational efficiency. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent retail method and system based on face recognition to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an intelligent retail method based on face recognition, applied to intelligent physical retail space;

[0007] Analyze shopping basket movement trajectories based on RFID tags on the shopping baskets, generate a dedicated interaction library based on the shopping basket movement trajectories, and identify high-frequency interaction hotspots. Use a data matching model to match purchased items with items in high-frequency interaction hotspots. Based on the matching results, identify abnormal behavior trajectories and eliminate data. Density estimation is performed on discrete movement trajectory data to generate a behavior heat map.

[0008] Based on the behavioral heat map, the retail space is divided into three types of retail areas: hot, warm, and cold. The hot, warm, and cold zones correspond to retail areas with gradually decreasing traffic. The hot zone is the retail area with the highest customer stay density and significant interaction frequency; the warm zone is the retail area with medium customer flow and transient dwell time; and the cold zone is the retail area with significantly low customer visit frequency and dwell time.

[0009] Based on the user's purchase data, the product association analysis is performed on the hot and cold areas of the retail space to generate a product association map;

[0010] Based on the product association graph, we analyze the strongly correlated product groups in the retail space, and reorganize the products in the hot and cold areas of the retail space through the strongly correlated product groups. This will avoid crowds and stampedes while promoting the sales of products in the cold area.

[0011] Furthermore, the specific method of generating a continuous behavior heat map based on the shopping basket movement trajectory includes:

[0012] The retail space is mapped to a coordinate system. Multiple mobile trajectories are tracked using RFID tags on shopping baskets to generate shopping basket movement trajectories and record the trajectory information of different users in the retail space: ut(x, y), where ut(x, y) represents the horizontal and vertical coordinate position information (x, y) of shopping basket u at the tth time node when it enters the retail space; u = 1, 2, 3...U, where U is the number of shopping baskets, and the horizontal and vertical coordinate points represent the location information of the goods shelves in the retail space respectively;

[0013] Using a spatial trajectory density visualization method, we estimate the density of discrete shopping basket movement trajectory data and generate a behavioral heat map. The frequency of shopping basket movement trajectory coverage is positively correlated with the color brightness of the behavior heat map. Specifically, the more times a retail space unit is covered by a trajectory, the higher the corresponding heat value, and the darker the color tone in the color scale mapping.

[0014] In order to avoid errors in the color scale display of the behavior heat map due to abnormal behavior, the discrete shopping basket movement trajectory data does not include abnormal behavior trajectory data, wherein the abnormal behavior trajectory data is determined, including:

[0015] Obtain any shopping basket trajectory u'-t(x, y). When there are consecutive time nodes where the shopping basket's location information remains unchanged, preliminarily determine that the user has interacted with the product, record the current shopping basket's location information, and generate a dedicated interaction library. Use an indicator function to count the number of occurrences of the same location information in the dedicated interaction library, and record the location information with the highest number of occurrences as a high-frequency interaction hotspot. When the maximum statistical count exceeds the total number of location information recorded in the shopping basket trajectory u'-t(x, y), it is determined that the high-frequency interaction hotspot contains potential user purchase behavior.

[0016] Based on the data matching model, it is determined whether the products purchased by the corresponding user in the shopping basket u' belong to the products in the high-frequency interaction hotspot area. When the products purchased by the user u' do not belong to the products in the high-frequency interaction hotspot area, the movement trajectory of any shopping basket is determined to be abnormal behavior trajectory data; otherwise, the movement trajectory is determined to be normal. The data matching model scans the products during the user payment process and matches the product data of the high-frequency interaction hotspot based on the product location database.

[0017] Furthermore, specific methods for conducting correlation analysis on products in hot and cold areas of retail space include:

[0018] Combining the products in the hot and cold areas of the retail space to generate sets R and L, R = {ri}, L = {lj}, ri is the i-th product in the hot area of ​​the retail space, lj is the j-th product in the cold area of ​​the retail space, ri and lj are the product numbers of the hot and cold areas of the retail space; i = 1, 2, 3...I, j = 1, 2, 3...J, I and J are both constants;

[0019] Filter the items that appear in both sets R and L in the historical purchase data of any user in a retail store to generate associated combinations (ri', lj'). Then, use collaborative filtering to count the number of times these associated combinations appear in the historical purchase data to obtain M(ri', lj'). The association value of the associated combination (ri', lj') of items in the hot and cold zones of the retail space is proportional to M(ri', lj'), where i'∈{1, 2, 3...I} and j'∈{1, 2, 3...J}.

[0020] A commodity association graph is constructed through the association combinations and the corresponding association values, wherein the commodities in the association combinations serve as nodes in the graph and the association values ​​serve as weights of the commodity connection links.

[0021] Furthermore, specific methods for reorganizing merchandise in hot and cold areas of a retail space include:

[0022] Select any product ri' in the hot zone, traverse the product ri' and the product lj' in the set L that have a connection link, and determine the association value between the product ri' and lj' through the product association graph, including: when there is only one connection link between the product ri' and lj', the association value of the two products is the weight of the link; when there are multiple connection links between the product ri' and lj', the weighted cumulative summation method is used to calculate the association value of the two products, specifically: there is a set of link products between the two products, and the corresponding link weight sequence is V = {v1, v2, v3...vN}, vN is the weight of the last connection link in the corresponding link weight sequence, N is the number of connection links between the product ri' and lj', then the association value between the product ri' and lj' is

[0023] When the correlation value between products ri' and lj' exceeds the preset correlation threshold, products ri' and lj' are set as a strongly correlated product group. By analyzing the replacement scores of hot zone products, the replacement method of the strongly correlated product group is determined:

[0024] According to the product association graph, determine the number of products ri' and lj' associated with sets L and R, as well as the corresponding association values. Calculate the replacement scores of the two products by taking the number of associated products, the sum of the association values ​​of the associated products, and the weighted product of the hot and cold area of ​​the two products.

[0025] When the replacement score of a cold zone product exceeds that of a hot zone product, product lj' is promoted to the hot zone corresponding to ri', which may stimulate the associated purchase potential of product lj' within a unit area; when the replacement score of a cold zone product is lower than that of a hot zone product, product ri' is copied to the cold zone corresponding to lj', which may drive the exposure intensity of associated products of cold zone ri' within a unit area; the strongly associated product groups in the retail store are traversed, and the products in the hot and cold areas of the retail space are reorganized according to the strongly associated product groups.

[0026] Furthermore, specific methods for dividing retail spaces into zones based on behavioral heat maps include:

[0027] The behavior heat map is initially divided into regions using an edge algorithm to determine the boundaries of different regions. The brightness value of each pixel in the region is weighted and accumulated to calculate the comprehensive activity of different areas of the retail space. The higher the pixel brightness value, the lower the comprehensive activity of different areas of the retail space. All areas are sorted from high to low according to the comprehensive activity, and the different areas of the retail space are divided into three zones: hot, warm, and cold using the quantile threshold method.

[0028] The pixel brightness value is obtained by decoding the pixels of the behavior heat map using a color space conversion algorithm, extracting the 8-bit quantized value (0-255) of the RGB (red, green, blue) channel of each pixel; and quantizing the depth change of the RGB value color through the lightness component in the HSV color model.

[0029] A smart retail system based on face recognition, comprising: a movement trajectory analysis module, a region division module, a commodity association analysis module, and a commodity reorganization module;

[0030] The mobile trajectory analysis module is used to track the movement trajectory of the shopping basket through RFID tags, and preliminarily determine whether the user interacts with the product based on whether the location information of the shopping basket changes at consecutive time nodes. When the user interacts with the product, the current location information of the shopping basket is recorded to generate a dedicated interaction library; the indicator function is used to count the number of times the same location information appears in the dedicated interaction library, and the location information with the highest number of appearances is recorded as a high-frequency interaction hotspot; based on the data matching model, the purchase trajectory of products that do not belong to the high-frequency interaction hotspot area is determined to be abnormal behavior trajectory data; and the spatial trajectory density visualization method is used to perform density estimation on the discrete shopping basket movement trajectory data of non-abnormal behavior trajectory data to generate a behavior heat map;

[0031] The region division module is used to perform a preliminary division of the behavior heat map using an edge algorithm to determine the boundaries of different regions, perform weighted accumulation based on the pixel brightness values ​​in each region, calculate the comprehensive activity of different regions of the retail space, and sort all regions from high to low according to the comprehensive activity. The different regions of the retail space are divided into three zones: hot, warm, and cold using a quantile threshold method;

[0032] The product association analysis module is used to generate sets by combing products in the hot and cold zones of the retail space, and to filter products that appear simultaneously in the product sets of the hot and cold zones in the historical product purchase data of any user in the retail store to generate associated combinations; to count the number of times the associated combinations appear in the historical product purchase data through collaborative filtering, wherein the association value between two products in the associated combination of products in the hot and cold zones of the retail space is proportional to the number of times the associated combination appears; and to construct a product association graph based on the associated combinations and their corresponding association values, wherein the products in the associated combinations serve as nodes in the graph, and the association values ​​serve as weights of the links connecting the products;

[0033] The commodity reorganization module is used to set strongly associated commodity groups when the association value between any commodity in the hot and cold zones exceeds a preset association threshold, traverse the strongly associated commodity groups in the retail store, and reorganize the commodities in the hot and cold zones in the retail space based on the strongly associated commodity groups.

[0034] Furthermore, the movement trajectory analysis module includes a movement trajectory tracking unit, an abnormal behavior trajectory data screening unit, and a behavior heat map generation unit;

[0035] The mobile trajectory tracking unit is used to implement multi-trajectory tracking through RFID tags and generate mobile trajectories; the abnormal behavior trajectory data screening unit is used to preliminarily determine whether the user interacts with the product based on whether the location information of the shopping basket at consecutive time nodes changes. When the user interacts with the product, the current shopping basket location information is recorded to generate an exclusive interaction library; the indicator function is used to count the number of times the same location information appears in the exclusive interaction library, and the location information with the highest number of appearances is recorded as a high-frequency interaction hotspot; based on the data matching model, the shopping basket movement trajectory is determined to be abnormal behavior trajectory data for purchased products that do not belong to the high-frequency interaction hotspot area; the behavior heat map generation unit is used to perform density estimation on the discrete shopping basket movement trajectory data of non-abnormal behavior trajectory data through a spatial trajectory density visualization method to generate a behavior heat map.

[0036] Furthermore, the commodity association analysis module includes an association combination analysis unit, a combination association value analysis unit, and a commodity association graph generation unit;

[0037] The association combination analysis unit is used to generate association combinations by screening the commodities that belong to sets R and L and appear simultaneously in the historical commodity purchase data of any user in the retail store; the combination association value analysis unit is used to count the number of times the association combination appears in the historical commodity purchase data through collaborative filtering, and the association value of two commodities in the commodity association combination of hot and cold zones in the retail space is proportional to the number of times the association combination appears; the commodity association graph generation unit is used to construct a commodity association graph using the association combinations and the corresponding association values, wherein the commodities in the association combination serve as nodes in the graph, and the association values ​​serve as weights of the commodity connection links.

[0038] Furthermore, the commodity reorganization module includes a commodity association value analysis unit, a commodity replacement evaluation unit, and a commodity reorganization unit;

[0039] The commodity association value analysis unit is used to determine the association value between any two commodities in the hot and cold zones through a commodity association graph, including: when there is only one connection link between the two commodities, the association value of the two commodities is the weight of the link; when there are multiple connection links between the two commodities, the association value of the two commodities is calculated by a weighted cumulative summation method; the commodity replacement evaluation unit is used to calculate the replacement score of the two commodities by the number of associated commodities corresponding to any two commodities in the hot and cold zones, the sum of the association values ​​of the associated commodities, and the weighted product of the hot and cold zone areas corresponding to the two commodities; when the replacement score of the cold zone commodity exceeds the replacement score of the hot zone commodity, the cold zone commodity is promoted to the hot zone corresponding to another commodity; when the replacement score of the cold zone commodity is lower than the replacement score of the hot zone commodity, the hot zone commodity is copied to the cold zone corresponding to another commodity; the commodity reorganization unit is used to traverse the strongly associated commodity groups in the retail store and reorganize the commodities in the hot and cold zones of the retail space according to the strongly associated commodity groups.

[0040] Compared with the existing technology, the beneficial effects achieved by the present invention are as follows: the present invention analyzes abnormal behavior trajectory data in the mobile trajectory through a data matching model to generate a behavior heat map; the retail space is divided into areas based on the behavior heat map, and the products are scientifically reorganized by the correlation values ​​of the products between the divided areas, thereby avoiding crowd congestion and significantly improving retail operation efficiency and customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0042] Figure 1 It is a flowchart of the steps of a scene design method based on virtual reality of the present invention. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0044] See also Figure 1 , the present invention provides a technical solution: the purpose of the present invention is to provide an intelligent retail method and system based on face recognition to solve the problems raised in the above background technology.

[0045] In order to solve the above technical problems, the present invention provides the following technical solution: This solution strictly complies with the relevant laws and regulations on personal information protection during implementation, and all account login processes involving facial recognition technology have obtained explicit authorization from the user. The specific process is as follows:

[0046] User consent required: When a user initiates an account login request through a mobile device or self-service device, the system will force the user agreement and privacy policy page to be displayed, clearly informing the user of the purpose of using facial recognition technology, the scope of data collection, and the storage period; the user must actively check the "I have read and agreed" option before entering the subsequent operation process. This step uses a double confirmation mechanism, that is, the user must slide to verify or enter a verification code to confirm the authenticity of the operation to avoid accidental authorization;

[0047] Face recognition technology starts with the user selecting "face login" as the authentication method to log in to their account.

[0048] After the user completes facial recognition login, he or she can receive a shopping basket equipped with an RFID (radio frequency identification) tag. The tag serves as a unique identity identifier and records in real time the movement trajectory of the shopping basket, length of stay, frequency of area visits, and other information during the user's shopping process.

[0049] A smart retail approach based on facial recognition, applied to smart physical retail spaces;

[0050] Analyze shopping basket movement trajectories based on RFID tags on the shopping baskets, generate a dedicated interaction library based on the shopping basket movement trajectories, and identify high-frequency interaction hotspots. Use a data matching model to match purchased items with items in high-frequency interaction hotspots. Based on the matching results, identify abnormal behavior trajectories and eliminate data. Density estimation is performed on discrete movement trajectory data to generate a behavior heat map.

[0051] Based on the behavioral heat map, the retail space is divided into three types of retail areas: hot, warm, and cold. The hot, warm, and cold zones correspond to retail areas with gradually decreasing traffic. The hot zone is the retail area with the highest customer stay density and significant interaction frequency; the warm zone is the retail area with medium customer flow and transient dwell time; and the cold zone is the retail area with significantly low customer visit frequency and dwell time.

[0052] Based on the user's purchase data, the product association analysis is performed on the hot and cold areas of the retail space to generate a product association map;

[0053] Based on the product association graph, we analyze the strongly correlated product groups in the retail space, and reorganize the products in the hot and cold areas of the retail space through the strongly correlated product groups. This will avoid crowds and stampedes while promoting the sales of products in the cold area.

[0054] Furthermore, the specific method of generating a continuous behavior heat map based on the shopping basket movement trajectory includes:

[0055] The retail space is mapped to a coordinate system. Multiple mobile trajectories are tracked using RFID tags on shopping baskets to generate shopping basket movement trajectories and record the trajectory information of different users in the retail space: ut(x, y), where ut(x, y) represents the horizontal and vertical coordinate position information (x, y) of shopping basket u at the tth time node when it enters the retail space; u = 1, 2, 3...U, where U is the number of shopping baskets, and the horizontal and vertical coordinate points represent the location information of the goods shelves in the retail space respectively;

[0056] Using a spatial trajectory density visualization method, we estimate the density of discrete shopping basket movement trajectory data and generate a behavioral heat map. The frequency of shopping basket movement trajectory coverage is positively correlated with the color brightness of the behavior heat map. Specifically, the more times a retail space unit is covered by a trajectory, the higher the corresponding heat value, and the darker the color tone in the color scale mapping.

[0057] In order to avoid errors in the color scale display of the behavior heat map due to abnormal behavior, the discrete shopping basket movement trajectory data does not include abnormal behavior trajectory data, wherein the abnormal behavior trajectory data is determined, including:

[0058] Obtain any shopping basket trajectory u'-t(x, y). When there are consecutive time nodes where the shopping basket's location information remains unchanged, preliminarily determine that the user has interacted with the product, record the current shopping basket's location information, and generate a dedicated interaction library. Use an indicator function to count the number of occurrences of the same location information in the dedicated interaction library, and record the location information with the highest number of occurrences as a high-frequency interaction hotspot. When the maximum statistical count exceeds the total number of location information recorded in the shopping basket trajectory u'-t(x, y), it is determined that the high-frequency interaction hotspot contains potential user purchase behavior.

[0059] Based on the data matching model, it is determined whether the products purchased by the corresponding user in the shopping basket u' belong to the products in the high-frequency interaction hotspot area. When the products purchased by the user u' do not belong to the products in the high-frequency interaction hotspot area, the movement trajectory of any shopping basket is determined to be abnormal behavior trajectory data; otherwise, the movement trajectory is determined to be normal. The data matching model scans the products during the user payment process and matches the product data of the high-frequency interaction hotspot based on the product location database.

[0060] Furthermore, specific methods for conducting correlation analysis on products in hot and cold areas of retail space include:

[0061] Combining the products in the hot and cold areas of the retail space to generate sets R and L, R = {ri}, L = {lj}, ri is the i-th product in the hot area of ​​the retail space, lj is the j-th product in the cold area of ​​the retail space, ri and lj are the product numbers of the hot and cold areas of the retail space; i = 1, 2, 3...I, j = 1, 2, 3...J, I and J are both constants;

[0062] Filter the items that appear in both sets R and L in the historical purchase data of any user in a retail store to generate associated combinations (ri', lj'). Then, use collaborative filtering to count the number of times these associated combinations appear in the historical purchase data to obtain M(ri', lj'). The association value of the associated combination (ri', lj') of items in the hot and cold zones of the retail space is proportional to M(ri', lj'), where i'∈{1, 2, 3...I} and j'∈{1, 2, 3...J}.

[0063] A commodity association graph is constructed through the association combinations and the corresponding association values, wherein the commodities in the association combinations serve as nodes in the graph and the association values ​​serve as weights of the commodity connection links.

[0064] Furthermore, specific methods for reorganizing merchandise in hot and cold areas of a retail space include:

[0065] Select any product ri' in the hot zone, traverse the product ri' and the product lj' in the set L that have a connection link, and determine the association value between the product ri' and lj' through the product association graph, including: when there is only one connection link between the product ri' and lj', the association value of the two products is the weight of the link; when there are multiple connection links between the product ri' and lj', the weighted cumulative summation method is used to calculate the association value of the two products, specifically: there is a set of link products between the two products, and the corresponding link weight sequence is V = {v1, v2, v3...vN}, vN is the weight of the last connection link in the corresponding link weight sequence, N is the number of connection links between the product ri' and lj', then the association value between the product ri' and lj' is

[0066] When the correlation value between products ri' and lj' exceeds the preset correlation threshold, products ri' and lj' are set as a strongly correlated product group. By analyzing the replacement scores of hot zone products, the replacement method of the strongly correlated product group is determined:

[0067] According to the product association graph, determine the number of products ri' and lj' associated with sets L and R, as well as the corresponding association values. Calculate the replacement scores of the two products by taking the number of associated products, the sum of the association values ​​of the associated products, and the weighted product of the hot and cold area of ​​the two products.

[0068] When the replacement score of a cold zone product exceeds that of a hot zone product, product lj' is promoted to the hot zone corresponding to ri', which may stimulate the associated purchase potential of product lj' within a unit area; when the replacement score of a cold zone product is lower than that of a hot zone product, product ri' is copied to the cold zone corresponding to lj', which may drive the exposure intensity of associated products of cold zone ri' within a unit area; the strongly associated product groups in the retail store are traversed, and the products in the hot and cold areas of the retail space are reorganized according to the strongly associated product groups.

[0069] Furthermore, specific methods for dividing retail spaces into zones based on behavioral heat maps include:

[0070] The behavior heat map is initially divided into regions using an edge algorithm to determine the boundaries of different regions. The brightness value of each pixel in the region is weighted and accumulated to calculate the comprehensive activity of different areas of the retail space. The higher the pixel brightness value, the lower the comprehensive activity of different areas of the retail space. All areas are sorted from high to low according to the comprehensive activity, and the different areas of the retail space are divided into three zones: hot, warm, and cold using the quantile threshold method.

[0071] The pixel brightness value is obtained by decoding the pixels of the behavior heat map using a color space conversion algorithm, extracting the 8-bit quantized value (0-255) of the RGB (red, green, blue) channel of each pixel; and quantizing the depth change of the RGB value color through the lightness component in the HSV color model.

[0072] A smart retail system based on face recognition, comprising: a movement trajectory analysis module, a region division module, a commodity association analysis module, and a commodity reorganization module;

[0073] The mobile trajectory analysis module is used to track the movement trajectory of the shopping basket through RFID tags, and preliminarily determine whether the user interacts with the product based on whether the location information of the shopping basket changes at consecutive time nodes. When the user interacts with the product, the current location information of the shopping basket is recorded to generate a dedicated interaction library; the indicator function is used to count the number of times the same location information appears in the dedicated interaction library, and the location information with the highest number of appearances is recorded as a high-frequency interaction hotspot; based on the data matching model, the purchase trajectory of products that do not belong to the high-frequency interaction hotspot area is determined to be abnormal behavior trajectory data; and the spatial trajectory density visualization method is used to perform density estimation on the discrete shopping basket movement trajectory data of non-abnormal behavior trajectory data to generate a behavior heat map;

[0074] The region division module is used to perform a preliminary division of the behavior heat map using an edge algorithm to determine the boundaries of different regions, perform weighted accumulation based on the pixel brightness values ​​in each region, calculate the comprehensive activity of different regions of the retail space, and sort all regions from high to low according to the comprehensive activity. The different regions of the retail space are divided into three zones: hot, warm, and cold using a quantile threshold method;

[0075] The product association analysis module is used to generate sets by combing products in the hot and cold zones of the retail space, and to filter products that appear simultaneously in the product sets of the hot and cold zones in the historical product purchase data of any user in the retail store to generate associated combinations; to count the number of times the associated combinations appear in the historical product purchase data through collaborative filtering, wherein the association value between two products in the associated combination of products in the hot and cold zones of the retail space is proportional to the number of times the associated combination appears; and to construct a product association graph based on the associated combinations and their corresponding association values, wherein the products in the associated combinations serve as nodes in the graph, and the association values ​​serve as weights of the links connecting the products;

[0076] The commodity reorganization module is used to set strongly associated commodity groups when the association value between any commodity in the hot and cold zones exceeds a preset association threshold, traverse the strongly associated commodity groups in the retail store, and reorganize the commodities in the hot and cold zones in the retail space based on the strongly associated commodity groups.

[0077] Furthermore, the movement trajectory analysis module includes a movement trajectory tracking unit, an abnormal behavior trajectory data screening unit, and a behavior heat map generation unit;

[0078] The mobile trajectory tracking unit is used to implement multi-trajectory tracking through RFID tags and generate mobile trajectories; the abnormal behavior trajectory data screening unit is used to preliminarily determine whether the user interacts with the product based on whether the location information of the shopping basket at consecutive time nodes changes. When the user interacts with the product, the current shopping basket location information is recorded to generate an exclusive interaction library; the indicator function is used to count the number of times the same location information appears in the exclusive interaction library, and the location information with the highest number of appearances is recorded as a high-frequency interaction hotspot; based on the data matching model, the shopping basket movement trajectory is determined to be abnormal behavior trajectory data for purchased products that do not belong to the high-frequency interaction hotspot area; the behavior heat map generation unit is used to perform density estimation on the discrete shopping basket movement trajectory data of non-abnormal behavior trajectory data through a spatial trajectory density visualization method to generate a behavior heat map.

[0079] Furthermore, the commodity association analysis module includes an association combination analysis unit, a combination association value analysis unit, and a commodity association graph generation unit;

[0080] The association combination analysis unit is used to generate association combinations by screening the commodities that belong to sets R and L and appear simultaneously in the historical commodity purchase data of any user in the retail store; the combination association value analysis unit is used to count the number of times the association combination appears in the historical commodity purchase data through collaborative filtering, and the association value of two commodities in the commodity association combination of hot and cold zones in the retail space is proportional to the number of times the association combination appears; the commodity association graph generation unit is used to construct a commodity association graph using the association combinations and the corresponding association values, wherein the commodities in the association combination serve as nodes in the graph, and the association values ​​serve as weights of the commodity connection links.

[0081] Furthermore, the commodity reorganization module includes a commodity association value analysis unit, a commodity replacement evaluation unit, and a commodity reorganization unit;

[0082] The commodity association value analysis unit is used to determine the association value between any two commodities in the hot and cold zones through a commodity association graph, including: when there is only one connection link between the two commodities, the association value of the two commodities is the weight of the link; when there are multiple connection links between the two commodities, the association value of the two commodities is calculated by a weighted cumulative summation method; the commodity replacement evaluation unit is used to calculate the replacement score of the two commodities by the number of associated commodities corresponding to any two commodities in the hot and cold zones, the sum of the association values ​​of the associated commodities, and the weighted product of the hot and cold zone areas corresponding to the two commodities; when the replacement score of the cold zone commodity exceeds the replacement score of the hot zone commodity, the cold zone commodity is promoted to the hot zone corresponding to another commodity; when the replacement score of the cold zone commodity is lower than the replacement score of the hot zone commodity, the hot zone commodity is copied to the cold zone corresponding to another commodity; the commodity reorganization unit is used to traverse the strongly associated commodity groups in the retail store and reorganize the commodities in the hot and cold zones of the retail space according to the strongly associated commodity groups.

[0083] In this embodiment:

[0084] Select any product ri' in the hot zone, traverse the product ri' and the products lj' in the set L that have a connection link, and determine the association value between the products ri' and lj' through the product association graph.

[0085] When there are 7 links between the products ri' and lj', the weighted cumulative summation method is used to calculate the association value of the two products. According to the existence of a set of link products between the two products, the corresponding link weight sequence is V = {1.4, 3.34, 1.65...}, and the association value between the products ri' and lj' is (∑N n = 1vn) / N = (∑N n = 1vn) / 7 = 0.76. The correlation value between products ri' and lj' exceeds the preset correlation threshold of 0.75. Products ri' and lj' are set as a strongly correlated product group. The replacement method for the strongly correlated product group is determined by analyzing the replacement scores of hot zone products: the number of products associated with products ri' and lj' in sets L and R and the corresponding correlation values ​​are determined based on the product association graph. The replacement scores of the two products are calculated by taking the number of associated products, the sum of the associated values ​​of the associated products, and the weighted product of the hot and cold zone areas of the two products. The replacement score of product ri' is 5 × (0.75 + 0.97 + ...) × 3 = 34.7, and the replacement score of product lj' is 7 × (0.35 + 0.67 + ...) × 10 = 108.

[0086] The replacement score of the cold zone product, 108, exceeds the replacement score of the hot zone product, 34.7, which promotes the product lj' to the hot zone corresponding to ri', and may stimulate the associated purchase potential of the product lj' within a unit area.

[0087] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0088] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A smart retail method based on facial recognition, applied to smart physical retail spaces, requires users to log in to their accounts and access their shopping carts through facial recognition, characterized by: Analyze shopping basket movement trajectories based on RFID tags on the shopping baskets, generate a dedicated interaction library based on the shopping basket movement trajectories, and identify high-frequency interaction hotspots. Use a data matching model to match purchased items with items in high-frequency interaction hotspots. Based on the matching results, identify abnormal behavior trajectories and eliminate data. Density estimation is performed on discrete movement trajectory data to generate a behavior heat map. Based on the behavior heat map, the retail space is divided into three retail areas: hot, warm, and cold areas. The hot, warm, and cold areas correspond to retail areas with gradually decreasing traffic; Based on the purchased product data, we conduct correlation analysis on the products in the hot and cold areas of the retail space and generate a product correlation map. Based on the product association graph, the strongly correlated product groups in the retail space are analyzed, and the products in the hot and cold areas of the retail space are reorganized through the strongly correlated product groups.

2. The smart retail method based on face recognition according to claim 1, characterized in that: The specific methods for generating continuous behavior heat maps based on shopping basket movement trajectories include: The retail space is mapped to a coordinate system, and multiple movement trajectories are tracked using RFID tags on shopping baskets to generate the shopping basket movement trajectory: ut(x, y), where ut(x, y) represents the horizontal and vertical coordinate position information (x, y) of shopping basket u at the tth time node when it enters the retail space; u = 1, 2, 3...U, where U is the number of shopping baskets, and the horizontal and vertical coordinate points represent the location information of the goods shelves in the retail space respectively; Using a spatial trajectory density visualization method, we estimate the density of discrete shopping basket movement trajectory data and generate a behavioral heat map. The frequency of shopping basket movement trajectory coverage is positively correlated with the color brightness of the behavior heat map. Specifically, the more times a retail space unit is covered by a trajectory, the higher the corresponding heat value, and the darker the color tone in the color scale mapping. The discrete shopping basket movement trajectory data does not include abnormal behavior trajectory data, wherein determining the abnormal behavior trajectory data includes: Obtain any shopping basket trajectory u'-t(x, y). When there are consecutive time nodes where the shopping basket's location information remains unchanged, preliminarily determine that the user has interacted with the product, record the current shopping basket's location information, and generate a dedicated interaction library. Use an indicator function to count the number of occurrences of the same location information in the dedicated interaction library, and record the location information with the highest number of occurrences as a high-frequency interaction hotspot. When the maximum statistical count exceeds the total number of location information recorded in the shopping basket trajectory u'-t(x, y), it is determined that the high-frequency interaction hotspot contains potential user purchase behavior. Based on the data matching model, it is determined whether the goods purchased by the user corresponding to the shopping basket u' belong to the goods in the high-frequency interaction hotspot area. When the purchased goods do not belong to the goods in the high-frequency interaction hotspot area, the movement trajectory of any shopping basket is determined to be abnormal behavior trajectory data; the data matching model scans the goods during the user payment process and matches the product data of the high-frequency interaction hotspot based on the product location database.

3. The smart retail method based on face recognition according to claim 1, characterized in that: Specific methods for correlation analysis of products in hot and cold areas of retail space include: Combining the products in the hot and cold areas of the retail space to generate sets R and L, R = {ri}, L = {lj}, ri is the i-th product in the hot area of ​​the retail space, lj is the j-th product in the cold area of ​​the retail space; i = 1, 2, 3...I, j = 1, 2, 3...J; Filter the items that appear in both sets R and L in the historical purchase data of any user in a retail store to generate associated combinations (ri', lj'). Then, use collaborative filtering to count the number of times these associated combinations appear in the historical purchase data to obtain M(ri', lj'). The association value of the associated combination (ri', lj') of items in the hot and cold zones of the retail space is proportional to M(ri', lj'), where i'∈{1, 2, 3...I} and j'∈{1, 2, 3...J}. A commodity association graph is constructed through the association combinations and the corresponding association values, wherein the commodities in the association combinations serve as nodes in the graph and the association values ​​serve as weights of the commodity connection links.

4. The smart retail method based on face recognition according to claim 3, characterized in that: Specific methods for reorganizing merchandise in hot and cold areas of a retail space include: Select any product ri' in the hot zone, traverse the product ri' and the products lj' in the set L that have a connection link, and determine the association value between products ri' and lj' through the product association graph. This includes: when there is only one connection link between products ri' and lj', the association value of the two products is the link weight; when there are multiple connection links between products ri' and lj', the weighted cumulative sum method is used to calculate the association value of the two products; When the correlation value between products ri' and lj' exceeds the preset correlation threshold, products ri' and lj' are set as a strongly correlated product group. The strongly correlated product groups in the retail store are traversed, and the products in the hot and cold areas of the retail space are reorganized based on the strongly correlated product groups.

5. The smart retail method based on face recognition according to claim 2, characterized in that: Specific methods for zoning retail spaces based on behavioral heat maps include: The behavior heat map is initially divided into regions using an edge algorithm to determine the boundaries of different regions. The brightness value of each pixel in the region is weighted and accumulated to calculate the comprehensive activity of different areas of the retail space. All areas are sorted from high to low according to the comprehensive activity, and the different areas of the retail space are divided into three zones: hot, warm, and cold using the quantile threshold method. The pixel brightness value is obtained by decoding the pixels of the behavior heat map using a color space conversion algorithm and extracting the 8-bit quantized value of the RGB channel of each pixel; The color depth of the RGB value is calculated by quantifying the lightness component in the HSV color model.

6. An intelligent retail system based on face recognition, characterized by: The smart retail system includes: a movement trajectory analysis module, a region division module, a commodity association analysis module and a commodity reorganization module; The mobile trajectory analysis module is used to track the movement trajectory of the shopping basket through RFID tags, and preliminarily determine whether the user interacts with the product based on whether the location information of the shopping basket changes at consecutive time nodes. When the user interacts with the product, the current location information of the shopping basket is recorded to generate a dedicated interaction library; the indicator function is used to count the number of times the same location information appears in the dedicated interaction library, and the location information with the highest number of appearances is recorded as a high-frequency interaction hotspot; based on the data matching model, the purchase trajectory of products that do not belong to the high-frequency interaction hotspot area is determined to be abnormal behavior trajectory data; and the spatial trajectory density visualization method is used to perform density estimation on the discrete shopping basket movement trajectory data of non-abnormal behavior trajectory data to generate a behavior heat map; The region division module is used to perform a preliminary division of the behavior heat map using an edge algorithm to determine the boundaries of different regions, perform weighted accumulation based on the pixel brightness values ​​in each region, calculate the comprehensive activity of different regions of the retail space, and sort all regions from high to low according to the comprehensive activity. The different regions of the retail space are divided into three zones: hot, warm, and cold using a quantile threshold method; The product association analysis module is used to generate sets by combing products in the hot and cold zones of the retail space, and to filter products that appear simultaneously in the product sets of the hot and cold zones in the historical product purchase data of any user in the retail store to generate associated combinations; to count the number of times the associated combinations appear in the historical product purchase data through collaborative filtering, wherein the association value between two products in the associated combination of products in the hot and cold zones of the retail space is proportional to the number of times the associated combination appears; and to construct a product association graph based on the associated combinations and their corresponding association values, wherein the products in the associated combinations serve as nodes in the graph, and the association values ​​serve as weights of the links connecting the products; The commodity reorganization module is used to set strongly associated commodity groups when the association value between any commodity in the hot and cold zones exceeds a preset association threshold, traverse the strongly associated commodity groups in the retail store, and reorganize the commodities in the hot and cold zones in the retail space based on the strongly associated commodity groups.

7. The face recognition-based smart retail system according to claim 6, characterized in that: The movement trajectory analysis module includes a movement trajectory tracking unit, an abnormal behavior trajectory data screening unit and a behavior heat map generation unit; The mobile trajectory tracking unit is used to implement multi-trajectory tracking through RFID tags to generate a mobile trajectory; the abnormal behavior trajectory data screening unit is used to preliminarily determine whether the user interacts with the product based on whether the location information of the shopping basket changes at consecutive time nodes. When the user interacts with the product, the current location information of the shopping basket is recorded to generate a dedicated interaction library; the indicator function is used to count the number of times the same location information appears in the dedicated interaction library, and the location information with the highest number of appearances is recorded as a high-frequency interaction hotspot; Based on the data matching model, the shopping basket movement trajectory is determined to be abnormal behavior trajectory data for purchased products that do not belong to the high-frequency interaction hotspot area; the behavior heat map generation unit is used to estimate the density of the discrete shopping basket movement trajectory data of the non-abnormal behavior trajectory data through a spatial trajectory density visualization method to generate a behavior heat map.

8. The face recognition-based smart retail system according to claim 7, characterized in that: The commodity association analysis module includes an association combination analysis unit, a combination association value analysis unit and a commodity association graph generation unit; The association combination analysis unit is used to generate association combinations by screening the commodities that belong to sets R and L and appear simultaneously in the historical commodity purchase data of any user in the retail store; the combination association value analysis unit is used to count the number of times the association combination appears in the historical commodity purchase data through collaborative filtering, and the association value of two commodities in the commodity association combination of hot and cold zones in the retail space is proportional to the number of times the association combination appears; the commodity association graph generation unit is used to construct a commodity association graph using the association combinations and the corresponding association values, wherein the commodities in the association combination serve as nodes in the graph, and the association values ​​serve as weights of the commodity connection links.

9. The face recognition-based smart retail system according to claim 8, characterized in that: The commodity reorganization module includes a commodity association value analysis unit and a commodity reorganization unit; The commodity association value analysis unit is used to determine the association value between any two hot and cold zones through a commodity association graph, including: when there is only one connection link between two commodities, the association value of the two commodities is the weight of the link; when there are multiple connection links between two commodities, the association value of the two commodities is calculated using a weighted cumulative summation method; the commodity reorganization unit is used to traverse the strongly associated commodity groups in the retail store and reorganize the commodities in the hot and cold zones of the retail space based on the strongly associated commodity groups.

Images