An e-commerce platform management method and related equipment suitable for children's shoes

Abstract

The present invention relates to the field of e-commerce technology, specifically to an e-commerce platform management method and related equipment suitable for children's shoes, comprising the following steps: extracting contour and curvature features through foot images, classifying foot types and evaluating arch fit, combining shoe structure data with user foot matching, calculating fit scores, evaluating inventory risk, calculating recommendation scores, adjusting display priorities, and generating a list of recommended products. In the present invention, by analyzing user foot images and extracting key features, combining arch curvature to achieve accurate shoe matching, evaluating the matching degree between the user's foot contour and shoe design, ensuring comfort and adaptability, and structural fit scores effectively reducing return and exchange issues, combining product warehousing time and storage cycle, identifying inventory risks, supporting timely inventory adjustments, reducing the backlog of unsaleable goods, optimizing inventory management, and improving platform operating efficiency and user shopping experience.

Description

Technical Field

[0001] The present invention relates to the field of e-commerce technology, and in particular to an e-commerce platform management method and related equipment suitable for children's shoes. Background Art

[0002] The field of e-commerce technology includes various methods and systems for realizing commodity transactions, service provision and information dissemination through computer networks. The core content of e-commerce technology is the optimization and management of transaction processes based on the network environment, including key links such as commodity display, user interaction, order processing, payment settlement, logistics tracking and after-sales service. E-commerce systems promote the digitalization, platformization and intelligent development of industries through the combination of data collection, information processing, customer behavior analysis and platform management strategies. They are widely used in multiple industries such as retail, manufacturing, cross-border trade, etc. With the increasing trend of consumer personalization and vertical segmentation of commodities, e-commerce platforms are gradually evolving towards professional and customized management models to adapt to the diverse characteristics of commodities and user needs.

[0003] Among them, the e-commerce platform management method suitable for children's shoes refers to the characteristics of complex category management, strong size adaptability and differentiated needs of parent users in the online sales of children's shoes products. By constructing product classification logic rules, configuring size recommendation models and establishing a user behavior label system, it can achieve structured management of product information, accurate product matching and personalized recommendation display. This method establishes a children's shoe attribute database based on data processing means, integrates product life cycle management specifications and inventory dynamic update mechanism, and combines the front-end page management interface configuration system to carry out product online management, user behavior record collection and product label classification, thereby assisting the platform to complete the setting of automated operation processes and category optimization.

[0004] Existing technologies mainly rely on product classification and size recommendations, but fail to deeply analyze the user's foot characteristics and the structural adaptability of shoes, resulting in insufficient response of the recommendation system to personalized needs. Returns and exchanges due to size or design mismatches often occur. Inventory management relies on static data updates and forecasts, lacks real-time monitoring of product life cycles and inventory risks, and is unable to adjust strategies in time to cope with changes in market demand, resulting in inventory backlogs or out-of-stock phenomena, reducing the platform's ability to operate accurately and user satisfaction, and making it difficult to meet increasingly personalized consumer needs. Summary of the Invention

[0005] In order to solve the technical problems existing in the prior art, the embodiment of the present invention provides an e-commerce platform management method and related equipment applicable to children's shoes. The technical solution is as follows:

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an e-commerce platform management method for children's shoes, comprising the following steps:

[0007] S1: Call multi-angle foot images, extract the edge pixels of the sole and dorsum of the foot, calculate the outer edge curvature change characteristics of multiple locations on the sole, use support vector machine to classify and identify the user's foot structure type, and output the foot type classification label;

[0008] S2: Calling the foot type classification label, extracting the arch contour curve from the side view of the foot to identify the arch structure characteristics, comparing it with the standard arch grade table to obtain the structure grade label, combining the sole rigidity score to identify the matching degree between the shoe model and the user's arch structure, and outputting the arch fit information;

[0009] S3: Retrieving the arch fit information, extracting the structural design data of the shoe model and comparing it with the user's foot contour features, evaluating the matching degree of each part, and calculating the shoe model structural fit score based on the arch fit;

[0010] S4: Extract the entry time of children's shoes products, calculate the storage days, and use the DBSCAN clustering algorithm to identify the inventory risk level of each product based on the storage cycle and sales trend data;

[0011] S5: Call the inventory risk level and the shoe structure adaptability score, through dynamic weight allocation and nonlinear coupling calculation, superimpose the storage cycle parameters and inventory turnover rate gradient, obtain the recommendation score value, and generate a recommended product list.

[0012] As a further solution of the present invention, the foot type classification label includes plantar contour type, instep contour type, and overall foot structure classification; the arch fit information includes arch curvature matching, arch structure level, and fit calculation results; the shoe structure fit score includes forefoot structure matching, heel structure matching, and instep space matching; the inventory risk level includes overdue risk level, short-term inventory level, and normal inventory level; the recommendation score includes shoe structure adaptation score, dynamic weight allocation list, and final recommendation priority.

[0013] As a further solution of the present invention, the steps of calling multi-angle foot images, extracting the edge pixels of the sole and dorsum contours, calculating the outer edge curvature change characteristics of multiple locations on the sole, and using a support vector machine to classify and identify the user's foot structure type and output a foot type classification label are as follows:

[0014] S101: Obtain multi-angle foot image data uploaded by the user, call the sole image and the dorsal image, use the image grayscale distribution detection method to extract the contour edge pixel sequence of the sole area and the dorsal area, and calculate the contour change rate at each position. Combined with the position information, a foot contour change rate sequence is generated;

[0015] S102: Calling the foot contour change rate sequence, dividing the sliding window into fixed-step windows, calculating curvature change characteristic values ​​at multiple locations, performing a second-order derivative operation on the change rate value within each window, identifying the outer edge curvature change characteristics of each location based on the change trend, and establishing a local feature value of the sole of the foot including a local change rate interval and a curvature change trend;

[0016] The outer edge curvature variation feature is calculated based on the second-order derivative of the foot contour change rate sequence within a fixed step-size sliding window, and is divided into high convex sections when the value is greater than 0.03 / mm², deep concave sections when the value is less than -0.03 / mm², and flat sections in between.

[0017] S103: Based on the local feature quantity of the sole, a support vector machine is used to identify and classify the category of each local feature of the user's foot, and the user's foot structure type label is output, including the foot width, foot length ratio and sole outer contour curvature label to obtain a foot type classification label.

[0018] As a further solution of the present invention, the steps of calling the foot type classification label, extracting the arch contour curve of the foot side image to identify the arch structure characteristics, comparing the standard arch grade table to obtain the structure grade label, combining the sole rigidity score to identify the matching degree between the shoe model and the user's arch structure, and outputting the arch fit information are specifically as follows:

[0019] S201: Calling the foot type classification label, obtaining a side image of the user's foot, extracting the boundary points of the arch area based on the Canny edge detection algorithm, collecting contour curve data in order of the boundary points, fitting the contour curve using the least squares method, calculating the curvature change value, and generating an arch curvature change sequence;

[0020] S202: Based on the arch curvature change sequence, curvature nodes are sampled at fixed intervals, the curvature change range and trend are analyzed, the curvature degree and structural extension characteristics are extracted, and the curvature interval benchmark value in the standard arch grade table is called to obtain the arch structure grade label;

[0021] The standard arch grading table calculates the average change in arch curvature and classifies a foot with a curvature greater than 0.045 / mm² as a high arch foot, a foot with a curvature between 0.025 / mm² and 0.045 / mm² as a normal foot, and a foot with a curvature less than 0.025 / mm² as a flat foot.

[0022] S203: Calculate the difference between the sole rigidity score of each shoe model and the center value of the target interval based on the arch structure grade label and the sole rigidity score of each shoe model, perform normalization processing, obtain the fit score, and establish arch fit information.

[0023] As a further solution of the present invention, the steps of calling the arch fit information, extracting the structural design data of the shoe model and comparing it with the user's foot contour features, evaluating the matching degree of each part, and calculating the shoe model structural fit score in combination with the arch fit are specifically as follows:

[0024] S301: Retrieving the arch fit information, extracting forefoot width, heel thickness, and instep space height parameters of each shoe model, matching the corresponding shoe model structural parameters based on the foot contour feature point data, calculating the size difference between each part, and obtaining the size difference values ​​between each shoe model and the user's foot at the forefoot, heel, and instep positions;

[0025] S302: Based on the part size difference values, a normalization method is used to normalize the multiple part difference values ​​to a unified range. Combined with the importance factor weight table of each part structure of the target shoe model, the average score of each standardized difference value and the corresponding structure importance weight is calculated to obtain the shoe structure matching score of the shoe model in the structural dimension;

[0026] The importance factor weight table assigns values ​​of 0.45, 0.35, and 0.20 to forefoot width, heel thickness, and instep space height, respectively, and normalizes the size difference of each part. The values ​​are then multiplied by the corresponding values ​​and summed to obtain the shoe structure matching score.

[0027] S303: Calculate the shoe structure fit score corresponding to each shoe style based on the shoe structure fit score and the arch fit information.

[0028] As a further solution of the present invention, the steps of extracting the storage time of children's shoes products, calculating the storage days, and using the DBSCAN clustering algorithm to identify the inventory risk level of each product based on the storage cycle and sales trend data are as follows:

[0029] S401: Obtain the storage time data of children's shoes products, calculate the difference between the storage time of each product and the current time, obtain the storage time information, and generate the storage time interval value;

[0030] S402: Based on the storage time interval value, by comparing it with a preset benchmark storage cycle interval level list, matching the storage cycle level label of each children's shoe product, and obtaining the storage cycle classification label corresponding to each product;

[0031] S403: calling the storage cycle classification label, analyzing the sales trend of each children's shoe product according to the sales data of each children's shoe product, and calculating the inventory risk level of each children's shoe product.

[0032] As a further solution of the present invention, the steps of calling the inventory risk level and the shoe structure fit score, performing dynamic weight allocation and nonlinear coupling calculation, superimposing the storage cycle parameter and the inventory turnover rate gradient, obtaining a recommendation score value, and generating a recommended product list are specifically as follows:

[0033] S501: Calling the inventory risk level and the shoe structure fit score, through dynamic weight allocation and nonlinear coupling calculation, superimposing the storage cycle parameter and the inventory turnover rate gradient, obtains the recommendation score value, and generates a product score list;

[0034] S502: Based on the product rating list, sort the children's shoes products according to the scores, extract the ranking number of each product, and adjust the display priority of each children's shoe product in the user recommendation interface to obtain the display priority number of each children's shoe product;

[0035] S503: Construct a product list for the user recommendation interface based on the display priority number, and map the basic information of each children's shoe product to the list, including the shoe name, shoe picture, sole rigidity score, and fit score, to obtain a recommended product list.

[0036] On the other hand, a device for managing an e-commerce platform for children's shoes is provided. The device is applied to a method for managing an e-commerce platform for children's shoes. The device includes:

[0037] The foot structure recognition module uses multi-angle foot image data uploaded by the user to extract the contour edge pixels in the sole and dorsal images. Based on the extracted curvature variation characteristics of the sole outer edge, it uses a support vector machine model to classify and identify the user's foot structure type, generate a foot type classification label, and pass it to the arch adaptation analysis module;

[0038] The arch fit analysis module calls the foot type classification label, extracts the arch contour curve from the side image of the foot, performs curve fitting, calculates the arch curvature change sequence, extracts the arch structural features, compares them with the standard arch grade table, obtains the arch structure grade label, combines the sole rigidity scores of multiple shoe models, evaluates the match between the shoe model and the user's arch structure, generates arch fit information, and transmits it to the shoe structure matching module;

[0039] The shoe structure matching module calls the arch fit information to extract the structural design data of each shoe model, including forefoot width, heel thickness, and instep space height parameters. It compares the user's foot contour feature data, evaluates the size matching of each structural part, and calculates the shoe structure fit score based on the arch fit. The generated shoe structure fit score data is then passed to the product recommendation generation module.

[0040] The inventory risk assessment module extracts the arrival time information of children's shoes products, calculates the number of days each product is stored, compares the storage time interval with a preset benchmark storage cycle interval level list, determines the storage cycle classification label, and combines sales trend information to obtain the inventory risk level of each children's shoe product. The generated inventory risk level data is then passed to the product recommendation generation module;

[0041] The product recommendation generation module calls the inventory risk level and the shoe structure adaptability score, and through dynamic weight allocation and nonlinear coupling calculation, superimposes the storage cycle parameters and inventory turnover rate gradient, dynamically adjusts the display priority of the products on the user recommendation interface according to the scoring results, and generates a recommended product list.

[0042] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0043] By analyzing user foot images and extracting key features, it is possible to achieve accurate shoe matching based on the foot structure type and arch curvature. By carefully comparing the user's foot contour with the shoe design data, the matching degree of each part is evaluated to ensure a high degree of fit in terms of comfort and adaptability. The structural fit scoring method combined with arch characteristics can effectively reduce the return and exchange problems caused by inappropriate shoes. With the help of analysis of the product entry time and storage cycle, the product inventory risk is identified, supporting the platform to make timely inventory adjustments, reduce the backlog of unsaleable goods, optimize inventory management, and improve the platform's operating efficiency and user shopping experience through precise matching and dynamic inventory adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0046] Figure 2 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0047] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0048] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0049] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0050] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0051] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0052] See also Figure 1 The present invention provides a technical solution, a method for managing an e-commerce platform for children's shoes, comprising the following steps:

[0053] S1: Call multi-angle foot images, extract the edge pixels of the sole and dorsum of the foot, calculate the outer edge curvature change characteristics of multiple locations on the sole, use support vector machine to classify and identify the user's foot structure type, and output the foot type classification label;

[0054] S2: Call the foot type classification label, extract the arch contour curve from the side view of the foot to identify the arch structure characteristics, compare it with the standard arch grade table to obtain the structure grade label, combine the sole rigidity score to identify the match between the shoe model and the user's arch structure, and output the arch fit information;

[0055] S3: Calling arch fit information, extracting the shoe's structural design data and comparing it with the user's foot contour features, evaluating the matching degree of each part, and combining the arch fit to calculate the shoe's structural fit score;

[0056] S4: Extract the entry time of children's shoes products, calculate the storage days, and use the DBSCAN clustering algorithm to identify the inventory risk level of each product based on the storage cycle and sales trend data;

[0057] S5: Call the inventory risk level and shoe structure adaptability score, through dynamic weight allocation and nonlinear coupling calculation, superimpose the storage cycle parameters and inventory turnover rate gradient, obtain the recommendation score value, and generate a list of recommended products.

[0058] The foot type classification labels include plantar contour type, instep contour type, and overall foot structure classification. The arch fit information includes arch curvature matching, arch structure level, and fit calculation results. The shoe structure fit score includes forefoot structure matching, heel structure matching, and instep space matching. The inventory risk level includes overdue risk level, short-term inventory level, and normal inventory level. The recommendation score includes shoe structure fit score, dynamic weight allocation list, and final recommendation priority.

[0059] See also Figure 1 , calling multi-angle foot images, extracting the edge pixels of the sole and dorsum of the foot, calculating the outer edge curvature change characteristics of multiple locations on the sole of the foot, using support vector machine to classify and identify the user's foot structure type, and outputting the foot type classification label. The specific steps include:

[0060] S101: Obtain multi-angle foot image data uploaded by the user, call the sole image and the dorsal image, use the image grayscale distribution detection method to extract the contour edge pixel sequence of the sole area and the dorsal area, and calculate the contour change rate at each position. Combined with the position information, a foot contour change rate sequence is generated;

[0061] To obtain multi-angle foot image data uploaded by users, all images are first standardized. The plantar and dorsal images are loaded at a fixed resolution of 500×500 pixels. For the plantar image, the grayscale value of each row of pixels is detected in turn, and points with grayscale values ​​between 80 and 180 are extracted and their corresponding row and column coordinates are recorded. For example, when scanning the 200th row, if the grayscale values ​​are in the allowed range for columns 150 to 160, these points are recorded as the preliminary plantar boundary. Subsequently, the dorsal image is subjected to a grayscale range of 90 to 190, and edge coordinates are extracted in the same way to complete the extraction of two sets of edge point sets. Afterwards, from the extracted point set, the sum of the squares of the differences between the horizontal and vertical coordinates of each two consecutive points is calculated and the square root is taken to obtain the contour change rate. For example, the change rate between points (120,220) and (122,223) is 2 pixels horizontally and 3 pixels vertically, so the change rate is about 3.606. After all points are processed in sequence, a change rate sequence is formed, for example , and combined with the position index to form complete foot contour change rate data for subsequent feature extraction;

[0062] Table 1 Grayscale range of sole and dorsum image extraction

[0063]

[0064] As shown in Table 1, by setting different grayscale ranges, the contour boundaries of the sole and dorsum of the foot can be accurately extracted, laying the foundation for the calculation of the change rate.

[0065] S102: Calling a foot contour change rate sequence, dividing the sliding window into fixed-step windows, calculating curvature change feature values ​​at multiple locations, performing a second-order derivative operation on the change rate value within each window, identifying the outer edge curvature change characteristics at each location based on the change trend, and establishing a local feature value of the plantar including a local change rate interval and a curvature change trend;

[0066] The specific formula for identifying the outer edge curvature change characteristics of each position based on the change trend is:

[0067] ;

[0068] Calculate the local curvature change eigenvalue;

[0069] in, Representative The local curvature variation eigenvalue within the position window, Representative The second-order difference value of the position change rate series, Representative The rate of change of position, Representative The rate of change of position, is the index position of the local sequence in the sliding window, is the current calculated position index, is the sliding window radius, For the The pixel coordinate value of the contour of the position, For the The pixel coordinate value of the contour of the position, is the standardized reference length parameter, To use the current index is the total number of samples contained in the center sliding window, Calculate the current position The index number that is offset one position forward, Calculate the current position The index number that is offset one position backward, For the The absolute value of the rate of change of position;

[0070] formula:

[0071] ;

[0072] Detailed explanation of the formula and the process of formula calculation and derivation:

[0073] The formula used to calculate the plantar contour in The local curvature change characteristic value of the position is used to identify the outer edge curvature change characteristics of the position, and then establish the local feature value of the plantar including the local change rate interval and curvature change trend;

[0074] Parameter meaning and setting value:

[0075] For the The second-order difference value of the position change rate series is calculated as follows: ,in For the The rate of change of position, For the The rate of change of position, For the Position change rate value, set , , ,but ;

[0076] For Center window radius The sum of the absolute values ​​of all change rates within the range, the sliding window radius Set to 5, set The values ​​are: 0.012, 0.014, 0.016, 0.018, 0.020, 0.022, 0.024, 0.026, 0.028, 0.030, 0.032, ;

[0077] is the total number of samples in the sliding window, substitute ,but ;

[0078] For the The pixel coordinate value of the contour of the position, For the The pixel coordinate value of the contour of the position, set , ,but ;

[0079] To standardize the reference length parameter, it is set to 10;

[0080] Substitute the parameters into the formula for calculation:

[0081] ;

[0082] ;

[0083] ;

[0084] ;

[0085] Substitute the above calculation results into the formula:

[0086] ;

[0087] result It shows the The local curvature change characteristic value within the position window is small, indicating that the plantar contour at this position changes relatively slowly. The result is used to identify the outer edge curvature change characteristics of this position, and then establish the local feature quantity of the plantar including the local change rate interval and curvature change trend.

[0088] The outer edge curvature variation feature is calculated based on the second-order derivative of the foot contour change rate sequence within a fixed step-length sliding window. It is divided into high-convex segments (greater than 0.03 / mm²), deep-concave segments (less than -0.03 / mm²), and flat segments in between.

[0089] S103: Using a support vector machine (SVM) to identify and classify each local feature of the user's foot based on the local feature quantity of the sole, output a user's foot structure type label, including foot width, foot length ratio, and sole outer contour curvature labels, to obtain a foot type classification label;

[0090] After extracting the local feature quantities of all windows, an encoding operation is performed on each feature quantity, and the change rate interval and smoothness trend are respectively assigned fixed digital codes. For example, the smooth interval is set to 0, the medium interval is set to 1, the violent interval is set to 2, the curvature change trend is set to 1 for increasing fluctuations, 2 for decreasing fluctuations, and 0 for stability. The feature quantities of each window form a set of codes. For example, if the change rate interval of a window is medium and the curvature change trend is increasing fluctuations, the code is All window codes are then integrated into a feature matrix as classification input to train a support vector machine model. The classification criteria are foot width (e.g., a foot width greater than 10cm is defined as wide), foot length ratio (foot length divided by foot width greater than 3.0 is defined as slender), and plantar curvature (a moderate change interval exceeding 60% is marked as moderate curvature). For example, if a user has a foot width of 10.8cm and a foot length of 32cm, and the change interval accounts for 70% of the medium change, the final classification label is wide, slender, and medium curvature as the output result.

[0091] Table 2 Foot local feature encoding rules

[0092]

[0093] As shown in Table 2, by clarifying the standard feature encoding, the uniformity and accuracy of the classification process can be ensured, which facilitates model training and prediction output.

[0094] See also Figure 1 , calling the foot type classification label, extracting the arch contour curve from the side view of the foot to identify the arch structure characteristics, comparing it with the standard arch grade table to obtain the structure grade label, combining the sole rigidity score to identify the matching degree between the shoe model and the user's arch structure, and outputting the arch fit information. The specific steps include:

[0095] S201: Calling the foot type classification label, obtaining a side image of the user's foot, extracting the boundary points of the arch area based on the Canny edge detection algorithm, collecting contour curve data in order of the boundary points, fitting the contour curve using the least squares method, calculating the curvature change value, and generating an arch curvature change sequence;

[0096] Retrieve and load the side image of the corresponding user's foot, unify the image size to 640×480 pixels, call the Canny edge detection operation, set the low threshold of 50 and the high threshold of 150, scan the image pixel by pixel, detect the edge contour points of the arch area, and filter out the continuously distributed and arc-shaped boundary point set based on the pixel position change pattern. For example, detect the edge point that changes continuously from the 100th row and 150th column to the 300th row and 170th column. Record the horizontal and vertical coordinates of each boundary point in the order of the edge points, and press Arrange in ascending order according to the row number to ensure the continuity of the contour curve, and then perform least square fitting on all contour points, that is, by minimizing the sum of the squares of the vertical distances from all points to the fitting curve, the optimal curve fitting parameters are obtained. After the fitting is completed, the curvature change is analyzed by the change in the tangent slope of each point on the fitting curve. The curvature change value is calculated by differential tangent angle change point by point. For example, if the tangent angles of two adjacent points are 10 degrees and 12 degrees respectively, the change is 2 degrees. Finally, the curvature change values ​​of all boundary points are organized into sequence data, such as curvature change sequence , output as a sequence of arch curvature changes;

[0097] Table 3 Example of arch area boundary point collection and curvature change

[0098]

[0099] As shown in Table 3, by extracting boundary points point by point and calculating the curvature change, a complete arch curvature change sequence is formed for subsequent arch characteristic recognition.

[0100] S202: Based on the arch curvature change sequence, curvature nodes are sampled at fixed intervals, the curvature change range and trend are analyzed, the curvature degree and structural extension characteristics are extracted, and the curvature interval benchmark value in the standard arch grade table is called to obtain the arch structure grade label;

[0101] Set 1 node to be sampled every 5 boundary points, such as in the curvature change sequence , the curvature values ​​of the 1st and 6th points are sampled in sequence to form a node curvature sequence , for the node curvature data after sampling, its variation range is counted, that is, the difference between the maximum curvature and the minimum curvature. For example, in the above node, the variation range is 0.03-0.03=0.00. The curvature variation trend is further judged by comparing the size relationship of the curvature values ​​of adjacent nodes. If the value of the latter node is greater than the value of the previous node, it is recorded as an increase, otherwise it is recorded as a decrease. Then, the curvature interval benchmark value specified in the standard arch grade table is called. Among them, the curvature mean greater than 0.045 / mm² is classified as high arch foot, the curvature mean between 0.025 / mm² and 0.045 / mm² is classified as normal foot, and the curvature mean less than 0.025 / mm² is classified as flat foot. For example, the actual sampling node curvature mean is 0.030 / mm², which belongs to the normal foot type. Finally, the corresponding arch structure grade label is output for subsequent shoe fit analysis.

[0102] Table 4 Standard table of arch curvature grades

[0103]

[0104] As shown in Table 4, the standardized classification based on the mean value of arch curvature change can clarify the structural level of the arch, which is convenient for subsequent fitness measurement;

[0105] The standard arch grade table calculates the average change in arch curvature and classifies a foot with a curvature greater than 0.045 / mm² as a high arch foot, a foot between 0.025 / mm² and 0.045 / mm² as a normal foot, and a foot with a curvature less than 0.025 / mm² as a flat foot.

[0106] S203: Calculating the difference between the sole stiffness score of each shoe model and the center value of the target interval based on the arch structure grade label and the sole stiffness score of each shoe model, performing normalization processing to obtain a fit score, and establishing arch fit information;

[0107] Based on the identified arch structure grade label, the sole stiffness score data of each shoe in the shoe database is retrieved. The target interval center value is determined for the user's grade. The center value is set to 0.8 for high arch feet, 0.5 for normal feet, and 0.3 for flat feet. For example, if the user is identified as a normal foot, the target interval center value is 0.5. Each shoe model is traversed and the sole stiffness score is subtracted from the target center value. The absolute value is taken to represent the preliminary difference. For example, if the sole stiffness score of a shoe model is 0.6, the preliminary difference is 0.1. To unify the difference comparison between different shoe models, all preliminary difference values ​​of all shoe models are normalized. Using the minimum-maximum normalization method, all difference values ​​are linearly scaled to the range of 0 to 1. Smaller normalized difference values ​​indicate higher fit. The final fit score is then obtained by subtracting the normalized difference value from 1. For example, if the normalized difference of a shoe model is 0.25, the fit is 0.75. Finally, all shoe models are ranked according to the fit score and a list of shoes with the highest fit is recommended to the user.

[0108] Table 5 Example of shoe sole rigidity and arch fit

[0109]

[0110] As shown in Table 5, by normalizing the difference between the sole stiffness score and the arch grade center value, the fit of different shoe models can be clarified, thereby making personalized recommendations.

[0111] See also Figure 1 , calling the arch fit information, extracting the shoe's structural design data and comparing it with the user's foot contour features, evaluating the matching degree of each part, and combining the arch fit to calculate the shoe's structural fit score. The specific steps include:

[0112] S301: Retrieving arch fit information, extracting forefoot width, heel thickness, and instep space height parameters for each shoe model, matching the corresponding shoe model structural parameters based on foot contour feature point data, calculating the size difference between each part, and obtaining the size difference values ​​between each shoe model and the user's foot at the forefoot, heel, and instep positions;

[0113] For each shoe model, its structural parameter data is extracted, including three parameters: forefoot width, heel thickness and instep space height. For example, the forefoot width of shoe model A is 98mm, the heel thickness is 28mm, and the instep space height is 65mm. At the same time, the user's foot contour feature point data is called, and the actual measured dimensions are extracted through the feature points of the forefoot area, heel area and the highest point area of ​​the instep. For example, the user's actual measured forefoot width is 102mm, the heel thickness is 30mm, and the instep space height is 68mm. The size of each part is matched separately, and the difference between the shoe parameters and the corresponding part parameters of the user is calculated. The calculation formula is the user size minus the shoe size, and the difference is taken as the absolute value. For example, the forefoot width difference is mm, the difference in heel thickness is mm, the height difference of the instep space is mm, obtain the size difference between shoe model A and the user in each part. All shoe models are calculated according to the same logic, and finally a list of size differences at the forefoot, heel, and instep of each shoe model is formed;

[0114] Table 6 Example of the difference between shoe models and user foot sizes

[0115]

[0116] As shown in Figure 6, by calculating the size differences between each part of the shoe and the user's foot structure item by item, basic data can be provided for subsequent structural matching analysis.

[0117] S302: Based on the part size difference values, a normalization method is used to standardize the multiple part difference values ​​to a unified range. Combined with the importance factor weight table of each part structure of the target shoe model, the average score of each standardized difference value and the corresponding structure importance weight is calculated to obtain the shoe structure matching score of the shoe model in the structural dimension;

[0118] Based on the previously calculated size differences, the differences in the forefoot, heel, and instep of each shoe model are first normalized. The normalization uses the minimum-maximum normalization method to linearly map the differences to the [0, 1] range. For example, for the forefoot differences of all shoe models, the maximum difference is 5mm and the minimum difference is 2mm. The normalized forefoot differences are calculated according to the following logic:

[0119] , taking the forefoot difference of shoe A as an example, the normalized calculation is , the same process is applied to the difference values ​​between the heel and the instep. Then, the structural importance factor weight table is called, and the forefoot width, heel thickness, and instep space height are weighted by 0.45, 0.35, and 0.20, respectively. The normalized difference value of each shoe model is multiplied by the corresponding weight value to obtain the weighted difference value of each part. The sum is then used to obtain the shoe structure matching score. For example, the normalized differences of shoe model A are 0.6667 for the forefoot, 0.3333 for the heel, and 0.5 for the instep. The calculation process is as follows:

[0120] Forefoot Score: ;

[0121] Score after heel: ;

[0122] Instep score: ;

[0123] The final structural matching score of shoe A is , similarly calculate the scores of all shoe models, which will be subsequently integrated with the arch fit;

[0124] The importance factor weight table assigns values ​​of 0.45, 0.35, and 0.20 to the forefoot width, heel thickness, and instep space height, respectively. After standardizing the size difference of each part, the values ​​are multiplied by the corresponding values ​​and the sum is calculated as the shoe structure matching score.

[0125] S303: Calculating the shoe structure fit score for each shoe style based on the shoe structure fit score and the arch fit information;

[0126] According to the structural matching score of each shoe model in the structural dimension, combined with the arch adaptation score of the corresponding shoe model, the shoe model structural adaptation score is comprehensively calculated. In the specific operation, the shoe model structural matching score and the arch adaptation score are first called separately, and the structural matching score of shoe model A is set to 0.5167 and the arch adaptation score is set to 0.75. In order to comprehensively consider the adaptability of the two aspects, the two are weighted averaged, where the arch adaptation weight is set to 0.6 and the structural matching weight is set to 0.4. The specific calculation steps are: the arch adaptation score of 0.75 is multiplied by 0.6 to obtain 0.45, and the structural matching score of 0.5167 is multiplied by 0.4 to obtain 0.2067. The two are added together to obtain a comprehensive adaptation score of 0.6567. The same method is used to calculate the remaining shoes, and the shoes are sorted by the comprehensive adaptation score to obtain a recommended list of shoes that best suit the user's foot shape and arch characteristics.

[0127] Table 7 Example of comprehensive shoe fit scores

[0128]

[0129] As shown in Table 7, by comprehensively calculating the matching of the arch and structure, the overall fit of the shoe can be accurately evaluated and personalized recommendations can be made.

[0130] See also Figure 1 , calling the arch fit information, extracting the shoe's structural design data and comparing it with the user's foot contour features, evaluating the matching degree of each part, and combining the arch fit to calculate the shoe's structural fit score. The specific steps include:

[0131] S401: Obtain the storage time data of children's shoes products, calculate the difference between the storage time of each product and the current time, obtain the storage time information, and generate the storage time interval value;

[0132] First, you need to query the entry record of each children's shoe product through the warehouse management system and read the "entry time" field in each product record. In this example, January 1, 2025 is used as the entry time of the example product A. Then call the system current time parameter and set it to April 29, 2025. Compare the entry time with the current time and perform time difference calculation. The current time stamp (such as April 29, 2025 is converted to a timestamp of 1751414400 seconds) is subtracted from the product entry timestamp (such as January 1, 2025, which is 1735689600 seconds). Then divide the difference by 86400 seconds (that is, the number of seconds in a day) to obtain the storage days. The calculation result is Days, rounded up to 18 days, perform the same steps for all products in batches to form a storage time information set, and further construct a storage time interval value list based on the storage days. For example, if product B is put into storage on December 15, 2024 (timestamp 1734211200 seconds), the storage days are Days, take 20 days, the storage time interval value is organized in array form as , days, etc., thereby completing the generation of the storage time interval value.

[0133] S402: Based on the storage time interval value, by comparing it with a preset benchmark storage cycle interval level list, matching the storage cycle level label of each children's shoe product, and obtaining the storage cycle classification label corresponding to each product;

[0134] Read the storage days obtained in the previous step in sequence. For each storage day, call the preset benchmark storage cycle interval level list for comparison. The list is set as follows:

[0135] Table 8 Warehousing cycle interval level table

[0136]

[0137] Using Table 8 as a reference, we read that product A has been stored for 18 days and judge that 18 is between 0 and 20, so it is matched to Class A. We read that product B has been stored for 20 days and judge that 20 is at the upper limit of Class A, so it is still classified as Class A. For the product with a storage period of 25 days, we judge that it is between 21 and 40 and match it to Class B. Similarly, for the product with a storage period of 45 days, we judge that 45 is between 41 and 60 and classify it as Class C. After matching the storage period of each product with the interval class, we generate the corresponding storage period classification label for each product, and finally form a list of corresponding children's shoes products to storage class labels, such as Product A: Class A, Product B: Class A, Product C: Class B, Product D: Class B, Product E: Class C.

[0138] S403: Calling the storage cycle classification label, analyzing the sales trend of each children's shoe product based on the sales data of each children's shoe product, and calculating the inventory risk level of each children's shoe product;

[0139] The specific formula for analyzing the sales trend of each children's shoe product is:

[0140] ;

[0141] Calculate the sales trend fluctuation value of each children's shoe product;

[0142] in, Representative The sales trend fluctuation value of children's shoes products, Representative Children's shoes products The number of sales within a time period, Representative The average sales volume of children's shoes products during the entire cycle, Representative Children's shoes products The number of days of storage cycle corresponding to the storage cycle classification label of a time period, Representative Children's shoes products Inventory levels within a time period, Representative The maximum inventory level of children's shoes during the analysis period, Representative The median sales volume of children's shoes products during the entire cycle, represents the total number of time periods used for analysis, Represents the number or index of each children's shoe product. An index representing a time period;

[0143] formula:

[0144] ;

[0145] Detailed explanation of the formula and the process of formula calculation and derivation:

[0146] This formula is used to calculate the sales trend fluctuation value of each children's shoe product. The results are used to assess the sales stability and inventory risk level of the product.

[0147] Parameter meaning and setting value:

[0148] For sales quantity, set : [100, 120, 110, 130, 140, 150, 160, 170, 180, 190, 200, 210];

[0149] is the storage cycle days, set to [30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30];

[0150] is the inventory level, set to : [50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160];

[0151] For the The maximum inventory level of children's shoes during the analysis period is set to ;

[0152] For the The median sales volume of children's shoes products in the entire cycle is set as ;

[0153] The total number of time periods used for analysis, set ;

[0154] Substitute the parameters into the formula for calculation:

[0155] Calculate average sales quantity :

[0156] ;

[0157] Calculate the absolute deviation of sales quantity from the average sales quantity :

[0158] ;

[0159] Calculation results:

[0160] ;

[0161] Calculate the square root of the storage cycle days :

[0162] ;

[0163] Calculate inventory reversal :

[0164] ;

[0165] Calculation results:

[0166] ;

[0167] Calculate the absolute deviation of the mean sales quantity from the median sales quantity :

[0168] ;

[0169] Substitute the above results into the formula to calculate :

[0170] ;

[0171] Calculation results:

[0172] ;

[0173] ;

[0174] ;

[0175] ;

[0176] ;

[0177] ;

[0178] The results 55 show that The sales trend fluctuation value of a children's shoe product is 55, indicating that the sales of this product fluctuated greatly during the analysis period, and there may be a risk of inventory backlog or unstable sales. Combining the moving average method and the exponential smoothing method, the sales curve of the product is fitted to predict future sales. Based on the predicted sales trend and current inventory level, taking into account factors such as seasonal fluctuations, supply cycles, and sales volatility of the product, the safety stock level is calculated. That is, to ensure that a certain sales capacity can still be maintained when the supply chain is disrupted. According to the inventory risk level assessment standard, products are divided into three levels: low risk, medium risk, and high risk. If the inventory level of a product is much higher than the safety stock level and the turnover rate is low, it can be classified as a high inventory risk level. Such products require special attention, including promotions or adjustments to sales strategies. Conversely, if the inventory level is close to or lower than the safety stock and the turnover rate is high, it is a low-risk product, indicating that the sales of the product are stable and the inventory is relatively sufficient.

[0179] See also Figure 1 The steps for generating a recommended product list include calling the inventory risk level and shoe structure fit score, and superimposing the storage cycle parameters and inventory turnover rate gradient through dynamic weight allocation and nonlinear coupling calculation to obtain the recommendation score.

[0180] S501: Call the inventory risk level and shoe structure fit score, through dynamic weight allocation and nonlinear coupling calculation, superimpose the storage cycle parameters and inventory turnover rate gradient, obtain the recommendation score value, and generate a product score list;

[0181] First, read the inventory risk level of each children's shoe product in the product database. Assuming that the risk level of product A is low risk and product B is medium risk, extract the shoe structure fit score. Assuming that the fit score of product A is 85 points and that of product B is 70 points, then perform dynamic weight allocation, set the inventory risk level weight coefficient to 0.6, and the shoe fit score weight coefficient to 0.4. When performing nonlinear coupling calculation, first multiply the two weights by the corresponding score values. For example, for product A: the inventory risk level score is assigned to 90, and the shoe fit score is 85. Execute 、 The sum of the two is 88. When the storage cycle parameter is superimposed, the storage days gradient is introduced as an adjustment factor. Assuming that the storage days gradient is assigned a value of 0.95 (the standard is 0.95 within 20 days, 0.9 for 21-40 days, 0.85 for 41-60 days, and 0.8 for more than 60 days, see Table 2 for the classification standard), commodity A has a storage period of 18 days, so the gradient is 0.95, and the final recommendation score is Similarly, for product B, the inventory risk score is 75 and the shoe fit score is 70. After weighting, we get 、 , the total score is 73, the storage days are 45 days, the corresponding gradient is 0.85, and the recommendation score is Finally, calculate the recommendation scores of all products in batches according to the above steps and generate a product score list;

[0182] Table 9 Storage cycle and storage time gradient table

[0183]

[0184] As shown in Table 9, different storage days ranges correspond to different gradients, which are used to adjust the recommendation score value.

[0185] S502: Based on the product rating list, sort the children's shoes products according to the scores, extract the ranking number of each product, and adjust the display priority of each children's shoe product in the user recommendation interface to obtain the display priority number of each children's shoe product;

[0186] First, sort the recommendation scores by score, and rank all children's shoe products from high to low. For example, Product A scores 83.6 and Product B scores 62.05, placing Product A first and Product B second. Extract the ranking number for each product. In this process, set the starting value of the number to 1 and number the products in descending order of score: Product A is numbered 1 and Product B is numbered 2. Then, adjust the display priority of each children's shoe product in the user recommendation interface. Use the product ranking number as the interface priority sequence identifier to update the number. For example, Product A is displayed as priority 1 on the interface, and Product B as priority 2. In the specific implementation, call the product display interface management module, read the recommended bit field, and replace the original field value with the currently obtained priority number. If there is a conflict in the original interface (such as an existing priority), compare the existing product score with the new product score, and the higher one will overwrite the lower one. After performing a priority refresh, save the update and finally obtain the display priority number set for each children's shoe product.

[0187] S503: Build a product list for the user recommendation interface based on the display priority number, and map the basic information of each children's shoe product to the list, including the shoe model name, shoe model image, sole rigidity score, and fit score, to obtain a recommended product list;

[0188] First, traverse the product data table in ascending order of display priority, and filter the corresponding product basic information fields. The field content includes shoe name, shoe picture, sole rigidity score and structural adaptability score. The sole rigidity score is quantified according to the preset standard, and the rigidity score range is set to 0-100 points. Assume that the sole rigidity score of product A is 78 points, and that of product B is 85 points. The adaptability score calls the adaptability score obtained in S501. Product A is 85 points, and product B is 70 points. The filtered shoe names such as "Qingbu Children's Shoes Type A" and "Jianbu Children's Shoes B" are added to the table. The shoe image paths are such as " / images / A.jpg" and " / images / B.jpg". The sole rigidity score and fit score are combined into a structured data list and mapped to the product list display module of the recommendation interface. In the display module, the list generation interface is called to write product entries in sequence according to priority. Serial number 1 corresponds to product A and serial number 2 corresponds to product B. Finally, a recommended product list is generated. Each row of the list displays the shoe model name, shoe image, sole rigidity score, and fit score, completing the data construction of the recommendation interface.

[0189] See also Figure 2 , an e-commerce platform management device for children's shoes, the e-commerce platform management device for children's shoes is used to execute the above-mentioned e-commerce platform management method for children's shoes, the device includes:

[0190] The foot structure recognition module uses multi-angle foot image data uploaded by the user to extract the contour edge pixels in the sole and dorsal images. Based on the extracted curvature variation characteristics of the sole outer edge, it uses a support vector machine model to classify and identify the user's foot structure type, generate a foot type classification label, and pass it to the arch adaptation analysis module;

[0191] The arch fit analysis module uses foot type classification labels to extract the arch contour curve from the side view of the foot, performs curve fitting, calculates the arch curvature change sequence, extracts arch structural features, compares them with a standard arch grade table, obtains an arch structure grade label, and combines the sole rigidity scores of multiple shoe models to evaluate the match between the shoe model and the user's arch structure. It then generates arch fit information and passes it to the shoe structure matching module.

[0192] The shoe structure matching module uses arch fit information to extract the structural design data of each shoe model, including forefoot width, heel thickness, and instep space height parameters. It then compares the user's foot contour feature data, evaluates the size matching of each structural part, and calculates the shoe structure fit score based on the arch fit. This score is then generated and passed to the product recommendation generation module.

[0193] The inventory risk assessment module extracts the arrival time information of children's shoes products, calculates the number of days each product is stored, compares the storage time interval with a preset benchmark storage cycle interval level list, determines the storage cycle classification label, and combines sales trend information to obtain the inventory risk level of each children's shoe product. The generated inventory risk level data is then passed to the product recommendation generation module;

[0194] The product recommendation generation module calls on the inventory risk level and shoe structure adaptability score, and through dynamic weight allocation and nonlinear coupling calculation, superimposes the storage cycle parameters and inventory turnover rate gradient, dynamically adjusts the display priority of the product in the user recommendation interface according to the scoring results, and generates a recommended product list.

[0195] The above embodiments can be implemented in whole or in part via software, hardware (e.g., circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0196] It should be understood that the term "and / or" as used herein simply describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the preceding and following related objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0197] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0198] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0199] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0200] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0201] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0202] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0203] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0204] If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical disks.

[0205] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for managing an e-commerce platform for children's shoes, characterized in that: The method comprises: S1: Call multi-angle foot images, extract the edge pixels of the sole and dorsum of the foot, calculate the outer edge curvature change characteristics of multiple locations on the sole, use support vector machine to classify and identify the user's foot structure type, and output the foot type classification label; S2: Calling the foot type classification label, extracting the arch contour curve from the side view of the foot to identify the arch structure characteristics, comparing it with the standard arch grade table to obtain the structure grade label, combining the sole rigidity score to identify the matching degree between the shoe model and the user's arch structure, and outputting the arch fit information; S3: Retrieving the arch fit information, extracting the structural design data of the shoe model and comparing it with the user's foot contour features, evaluating the matching degree of each part, and calculating the shoe model structural fit score based on the arch fit; S4: Extract the entry time of children's shoes products, calculate the storage days, and use the DBSCAN clustering algorithm to identify the inventory risk level of each product based on the storage cycle and sales trend data; S5: Calling the inventory risk level and the shoe structure fit score, through dynamic weight allocation and nonlinear coupling calculation, superimposing the storage cycle parameter and the inventory turnover rate gradient, obtains the recommendation score value, and generates a recommended product list; The steps for calling multi-angle foot images, extracting the edge pixels of the sole and dorsum contours, calculating the outer edge curvature change characteristics at multiple locations on the sole, and using a support vector machine to classify and identify the user's foot structure type and output the foot type classification label are as follows: S101: Obtain multi-angle foot image data uploaded by the user, call the sole image and the dorsal image, use the image grayscale distribution detection method to extract the contour edge pixel sequence of the sole area and the dorsal area, and calculate the contour change rate at each position. Combined with the position information, a foot contour change rate sequence is generated; S102: Calling the foot contour change rate sequence, dividing the sliding window into fixed-step windows, calculating curvature change characteristic values ​​at multiple locations, performing a second-order derivative operation on the change rate value within each window, identifying the outer edge curvature change characteristics of each location based on the change trend, and establishing a local feature value of the sole of the foot including a local change rate interval and a curvature change trend; The outer edge curvature variation feature is calculated based on the second-order derivative of the foot contour change rate sequence within a fixed step-size sliding window, and is divided into high convex sections when the value is greater than 0.03 / mm², deep concave sections when the value is less than -0.03 / mm², and flat sections in between. S103: Based on the local feature quantity of the sole, a support vector machine is used to identify and classify the category of each local feature of the user's foot, and the user's foot structure type label is output, including the foot width, foot length ratio and sole outer contour curvature label to obtain a foot type classification label.

2. The e-commerce platform management method for children's shoes according to claim 1, characterized in that: The foot type classification label includes the plantar contour type, the instep contour type, and the overall foot structure classification; the arch fit information includes the arch curvature matching, the arch structure level, and the fit calculation result; the shoe structure fit score includes the forefoot structure matching, the heel structure matching, and the instep space matching; the inventory risk level includes the overdue risk level, the short-term inventory level, and the normal inventory level; the recommendation score includes the shoe structure fit score, the dynamic weight allocation list, and the final recommendation priority.

3. The e-commerce platform management method for children's shoes according to claim 1, characterized in that: The specific formula for identifying the outer edge curvature change characteristics of each position based on the change trend is: ； Calculate the local curvature change eigenvalue; in, Representative The local curvature variation eigenvalue within the position window, Representative The second-order difference value of the position change rate series, Representative The rate of change of position, Representative The rate of change of position, is the index position of the local sequence in the sliding window, is the current calculated position index, is the sliding window radius, For the The pixel coordinate value of the contour of the position, For the The pixel coordinate value of the contour of the position, is the standardized reference length parameter, To use the current index is the total number of samples contained in the center sliding window, Calculate the current position The index number that is offset one position forward, Calculate the current position The index number that is offset one position backward, For the The absolute value of the rate of change of position.

4. The e-commerce platform management method for children's shoes according to claim 3 is characterized in that: The steps of calling the foot type classification label, extracting the arch contour curve from the side view of the foot to identify the arch structure characteristics, comparing the standard arch grade table to obtain the structure grade label, combining the sole rigidity score to identify the matching degree between the shoe model and the user's arch structure, and outputting the arch fit information are as follows: S201: Calling the foot type classification label, obtaining a side image of the user's foot, extracting the boundary points of the arch area based on the Canny edge detection algorithm, collecting contour curve data in order of the boundary points, fitting the contour curve using the least squares method, calculating the curvature change value, and generating an arch curvature change sequence; S202: Based on the arch curvature change sequence, curvature nodes are sampled at fixed intervals, the curvature change range and trend are analyzed, the curvature degree and structural extension characteristics are extracted, and the curvature interval benchmark value in the standard arch grade table is called to obtain the arch structure grade label; The standard arch grading table calculates the average change in arch curvature and classifies a foot with a curvature greater than 0.045 / mm² as a high arch foot, a foot with a curvature between 0.025 / mm² and 0.045 / mm² as a normal foot, and a foot with a curvature less than 0.025 / mm² as a flat foot. S203: Calculate the difference between the sole rigidity score of each shoe model and the center value of the target interval based on the arch structure grade label and the sole rigidity score of each shoe model, perform normalization processing, obtain the fit score, and establish arch fit information.

5. The e-commerce platform management method for children's shoes according to claim 4 is characterized in that: The steps of calling the arch fit information, extracting the shoe's structural design data, comparing it with the user's foot contour features, evaluating the matching degree of each part, and calculating the shoe's structural fit score based on the arch fit are as follows: S301: Retrieving the arch fit information, extracting forefoot width, heel thickness, and instep space height parameters of each shoe model, matching the corresponding shoe model structural parameters based on the foot contour feature point data, calculating the size difference between each part, and obtaining the size difference values ​​between each shoe model and the user's foot at the forefoot, heel, and instep positions; S302: Based on the part size difference values, a normalization method is used to normalize the multiple part difference values ​​to a unified range. Combined with the importance factor weight table of each part structure of the target shoe model, the average score of each standardized difference value and the corresponding structure importance weight is calculated to obtain the shoe structure matching score of the shoe model in the structural dimension; The importance factor weight table assigns values ​​of 0.45, 0.35, and 0.20 to forefoot width, heel thickness, and instep space height, respectively, and normalizes the size difference of each part. The values ​​are then multiplied by the corresponding values ​​and summed to obtain the shoe structure matching score. S303: Calculate the shoe structure fit score corresponding to each shoe style based on the shoe structure fit score and the arch fit information.

6. The e-commerce platform management method for children's shoes according to claim 5, characterized in that: The steps for extracting the arrival time of children's shoes, calculating the storage days, and using the DBSCAN clustering algorithm to identify the inventory risk level of each product based on the storage cycle and sales trend data are as follows: S401: Obtain the storage time data of children's shoes products, calculate the difference between the storage time of each product and the current time, obtain the storage time information, and generate the storage time interval value; S402: Based on the storage time interval value, by comparing it with a preset benchmark storage cycle interval level list, matching the storage cycle level label of each children's shoe product, and obtaining the storage cycle classification label corresponding to each product; S403: calling the storage cycle classification label, analyzing the sales trend of each children's shoe product according to the sales data of each children's shoe product, and calculating the inventory risk level of each children's shoe product.

7. The e-commerce platform management method for children's shoes according to claim 6, characterized in that: The specific formula for analyzing the sales trend of each children's shoe product is: ； Calculate the sales trend fluctuation value of each children's shoe product; in, Representative The sales trend fluctuation value of children's shoes products, Representative Children's shoes products The number of sales within a time period, Representative The average sales volume of children's shoes products during the entire cycle, Representative Children's shoes products The number of days of storage cycle corresponding to the storage cycle classification label of a time period, Representative Children's shoes products Inventory levels within a time period, Representative The maximum inventory level of children's shoes during the analysis period, Representative The median sales volume of children's shoes products during the entire cycle, represents the total number of time periods used for analysis, Represents the number or index of each children's shoe product. An index representing a time period.

8. The e-commerce platform management method for children's shoes according to claim 7, characterized in that: The steps of calling the inventory risk level and the shoe structure fit score, performing dynamic weight allocation and nonlinear coupling calculation, and superimposing the storage cycle parameter and inventory turnover gradient to obtain the recommendation score and generate a recommended product list are as follows: S501: Calling the inventory risk level and the shoe structure fit score, through dynamic weight allocation and nonlinear coupling calculation, superimposing the storage cycle parameter and the inventory turnover rate gradient, obtains the recommendation score value, and generates a product score list; S502: Based on the product rating list, sort the children's shoes products according to the scores, extract the ranking number of each product, and adjust the display priority of each children's shoe product in the user recommendation interface to obtain the display priority number of each children's shoe product; S503: Construct a product list for the user recommendation interface based on the display priority number, and map the basic information of each children's shoe product to the list, including the shoe name, shoe picture, sole rigidity score, and fit score, to obtain a recommended product list.

9. An e-commerce platform management device suitable for children's shoes, characterized by: The device is used to implement the e-commerce platform management method for children's shoes according to any one of claims 1 to 8, and the device includes: The foot structure recognition module uses multi-angle foot image data uploaded by the user to extract the contour edge pixels in the sole and dorsal images. Based on the extracted curvature variation characteristics of the sole outer edge, it uses a support vector machine model to classify and identify the user's foot structure type, generate a foot type classification label, and pass it to the arch adaptation analysis module; The arch fit analysis module calls the foot type classification label, extracts the arch contour curve from the side image of the foot, performs curve fitting, calculates the arch curvature change sequence, extracts the arch structural features, compares them with the standard arch grade table, obtains the arch structure grade label, combines the sole rigidity scores of multiple shoe models, evaluates the match between the shoe model and the user's arch structure, generates arch fit information, and transmits it to the shoe structure matching module; The shoe structure matching module calls the arch fit information to extract the structural design data of each shoe model, including forefoot width, heel thickness, and instep space height parameters. It compares the user's foot contour feature data, evaluates the size matching of each structural part, and calculates the shoe structure fit score based on the arch fit. The generated shoe structure fit score data is then passed to the product recommendation generation module. The inventory risk assessment module extracts the arrival time information of children's shoes products, calculates the number of days each product is stored, compares the storage time interval with a preset benchmark storage cycle interval level list, determines the storage cycle classification label, and combines sales trend information to obtain the inventory risk level of each children's shoe product. The generated inventory risk level data is then passed to the product recommendation generation module; The product recommendation generation module calls the inventory risk level and the shoe structure adaptability score, and through dynamic weight allocation and nonlinear coupling calculation, superimposes the storage cycle parameters and inventory turnover rate gradient, dynamically adjusts the display priority of the products on the user recommendation interface according to the scoring results, and generates a recommended product list.

Images