Virtual three-dimensional fitting system for outdoor jacket

Abstract

The invention discloses an outdoor jacket virtual three-dimensional fitting system, and belongs to the technical field of outdoor jacket virtual fitting, and the system comprises a three-dimensional model construction unit which is configured to collect and process outdoor jacket real-time point cloud data and construct an outdoor jacket virtual three-dimensional model; the human body model matching unit is configured to accurately match the outdoor jacket virtual three-dimensional model with human body measurement data; the virtual three-dimensional fitting unit is configured for a user to carry out virtual reality three-dimensional fitting experience on the outdoor jacket, simulate the texture of the jacket and fitting actions and check the dynamic effect of the outdoor jacket; and the automatic size recommendation unit is configured to automatically recommend the optimal virtual reality three-dimensional fitting size of the outdoor jacket to the user. According to the invention, the problems of poor shopping experience and high return rate of consumers caused by incapability of enabling users to perform fitting experience in a virtual environment in the prior art are solved. According to the invention, the user can perform fitting experience in the virtual environment, the shopping experience of consumers can be effectively improved, and the return rate is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual fitting of windbreakers, and specifically provides a virtual three-dimensional fitting system for windbreakers. Background Art

[0002] In the field of clothing consumption, people increasingly favor the "convenient shopping" mode, and the most representative one is the consumption mode of "online shopping". Compared with shopping in a specific shopping mall, the greatest advantage of online clothing shopping is convenience, speed, time-saving and labor-saving. However, the consumption method of staying at home also brings new problems. Since it is difficult to contact the actual clothing before purchase, there is a large deviation between the final wearing effect of the customer and their expectation. High negative review rate and high return rate are also the biggest problems brought by the consumption mode of online clothing shopping.

[0003] Chinese Patent with Publication No. CN119520704A discloses a virtual fitting mirror system, including a base, on the upper side of which a mirror frame is fixedly connected. Inside the mirror frame, a one-way perspective glass and a display screen are fixedly connected. The one-way perspective glass and the display screen respectively occupy half of the inside of the mirror frame. On the back side of the one-way perspective glass, an installation box is fixedly connected. Inside the installation box, a camera is fixedly connected. The front side of the camera is a reflective mirror surface, and the back side of the camera is a perspective mirror surface. The camera can take pictures of the front side through the back side of the one-way perspective glass. Users can conveniently view different colors and patterns of the same clothing style without having to try on and change multiple times, which is not only convenient and fast, but also helps to improve the satisfaction with the fitting effect. However, this patent has the following defects:

[0004] The existing technology cannot enable users to have a fitting experience in a virtual environment, resulting in a poor shopping experience for consumers and a high return rate. Summary of the Invention

[0005] The purpose of the present invention is to provide a virtual three-dimensional fitting system for windbreakers, which allows users to have a fitting experience in a virtual environment, can effectively improve the shopping experience of consumers, reduce the return rate, and solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] A virtual three-dimensional fitting system for windbreakers, comprising:

[0008] A three-dimensional model construction unit configured to collect real-time point cloud data of a windbreaker, process it to determine the characteristic point cloud data of the windbreaker, correlate the characteristic point cloud data of the windbreaker according to the cosine similarity value corresponding to each characteristic point cloud data of the windbreaker and its maximum and minimum values on the main axis, determine the associated set of the characteristic point cloud data of the windbreaker, and construct a virtual three-dimensional model of the windbreaker;

[0009] A mannequin matching unit, configured to precisely match the virtual 3D model of the windbreaker with the body measurement data, so that the user's body fits the windbreaker;

[0010] A virtual 3D fitting unit, configured to enable the user to experience virtual reality 3D fitting of the windbreaker, simulate the texture of the clothing and the fitting actions, and view the dynamic effect of the windbreaker.

[0011] Preferably, real-time point cloud data of the windbreaker is collected, including:

[0012] Based on 3D scanning technology, a 3D laser scanner is used to perform real-time scanning and collection of the style of the windbreaker, and the style point cloud data of the windbreaker is obtained;

[0013] Based on 3D scanning technology, a 3D laser scanner is used to perform real-time scanning and collection of the texture of the windbreaker, and the texture point cloud data of the windbreaker is obtained;

[0014] Based on 3D scanning technology, a 3D laser scanner is used to perform real-time scanning and collection of the fabric of the windbreaker, and the fabric point cloud data of the windbreaker is obtained;

[0015] Among them, according to the style point cloud data, texture point cloud data and fabric point cloud data of the windbreaker, the real-time point cloud data of the windbreaker is determined.

[0016] Preferably, the real-time point cloud data of the windbreaker is processed, including:

[0017] Clean the real-time point cloud data of the windbreaker to remove the noise data that is useless for the virtual 3D fitting of the windbreaker in the real-time point cloud data of the windbreaker;

[0018] Check and identify the real-time point cloud data of the windbreaker, determine whether there are duplicate values, missing values and abnormal values that are useless for the virtual 3D fitting of the windbreaker in the real-time point cloud data of the windbreaker, and delete the duplicate values, missing values and abnormal values that are useless for the virtual 3D fitting of the windbreaker in the real-time point cloud data of the windbreaker;

[0019] Extract features from the real-time point cloud data of the windbreaker, extract the features useful for the virtual 3D fitting of the windbreaker from the real-time point cloud data of the windbreaker, and determine the feature point cloud data of the windbreaker.

[0020] Preferably, a virtual 3D model of the windbreaker is constructed, including:

[0021] Associate the feature point cloud data of the windbreaker to determine the associated set of the feature point cloud data of the windbreaker;

[0022] Based on the virtual 3D fitting requirements of windbreakers, according to the associated set of characteristic point cloud data of windbreakers, select 3D modeling software. Based on the selected 3D modeling software, perform 3D modeling on the associated set of characteristic point cloud data of windbreakers to construct a 1:1 restored virtual 3D model of the windbreaker;

[0023] Perform detail carving and texture rendering on the virtual 3D model of the windbreaker. Use carving tools to add wrinkles, stitches, and texture details, assign fabric materials to different parts of the windbreaker, and set roughness, metallicity, and normal parameters, and render the texture, accurately mapping it to the model surface to determine the virtual 3D model of the windbreaker based on virtual reality.

[0024] Preferably, associate the characteristic point cloud data of the windbreaker to determine the associated set of characteristic point cloud data of the windbreaker, including:

[0025] Scan all the characteristic point cloud data of the windbreaker and extract the characteristic data of each characteristic point cloud data of the windbreaker;

[0026] Use the characteristic data to obtain the cosine similarity value between every two characteristic point cloud data among all the characteristic point cloud data of the windbreaker;

[0027] Compare the cosine similarity value between every two characteristic point cloud data of the windbreaker with a preset similarity threshold;

[0028] Filter out the characteristic point cloud data of the windbreaker with a cosine similarity value less than the preset similarity threshold to form a target point cloud data set;

[0029] For each characteristic point cloud data in the target point cloud data set, obtain the projection value of each characteristic point cloud data of the windbreaker projected onto the main axis of its corresponding oriented bounding box (OBB);

[0030] Obtain the main axis components corresponding to the three-axis data of each characteristic point cloud data of the windbreaker from the projection values;

[0031] Use the main axis components to obtain the maximum and minimum values of each characteristic point cloud data of the windbreaker on the main axis;

[0032] Associate the characteristic point cloud data of the windbreaker according to the cosine similarity value corresponding to each characteristic point cloud data of the windbreaker and its maximum and minimum values on the main axis to determine the associated set of characteristic point cloud data of the windbreaker.

[0033] Preferably, associating the characteristic point cloud data of the windbreaker according to the cosine similarity value corresponding to each characteristic point cloud data of the windbreaker and its maximum and minimum values on the main axis includes:

[0034] Extract the cosine similarity value corresponding to each feature point cloud data of the windbreaker and its maximum and minimum values on the main axis;

[0035] Combine the cosine similarity value corresponding to each feature point cloud data of the windbreaker and its maximum and minimum values on the main axis into a dual-domain feature vector;

[0036] Among them, the dual-domain feature vector is obtained through the following formula:

[0037]

[0038] Among them, D represents the dual-domain feature vector corresponding to each feature point cloud data of the windbreaker; S represents the cosine similarity value corresponding to each feature point cloud data of the windbreaker; X max and X min respectively represent the maximum and minimum values on the main axis corresponding to each feature point cloud data of the windbreaker; a represents the base of the logarithmic function, and the base is determined according to actual application requirements;

[0039] Construct a three-dimensional projection tensor corresponding to each feature point cloud data of the windbreaker for each feature point cloud data of the windbreaker;

[0040] Among them, the three-dimensional projection tensor is obtained through the following formula:

[0041]

[0042] Among them, H represents the three-dimensional projection tensor corresponding to each feature point cloud data of the windbreaker; K x 、K y and K z respectively represent the geometric spans of each feature point cloud data of the windbreaker on the x-axis, y-axis, and z-axis; S represents the cosine similarity value corresponding to each feature point cloud data of the windbreaker;

[0043] Generate the correlation degree between every two feature point cloud data of the windbreaker by combining the three-dimensional projection tensor with the dual-domain feature vector;

[0044] Among them, the correlation degree between every two feature point cloud data of the windbreaker is obtained through the following formula:

[0045]

[0046] Among them, G ij represents the correlation degree between the i-th feature point cloud data of the windbreaker and the j-th feature point cloud data of the windbreaker; D i and H i respectively represent the dual-domain feature vector and the three-dimensional projection tensor corresponding to the i-th feature point cloud data of the windbreaker; D j and H jrespectively represent the dual-domain feature vector and the three-dimensional projection tensor corresponding to the j-th windbreaker feature point cloud data; X maxi and X mini respectively represent the maximum and minimum values on the main axis corresponding to the i-th windbreaker feature point cloud data; X maxj and X minj respectively represent the maximum and minimum values on the main axis corresponding to the j-th windbreaker feature point cloud data;

[0047] Compare the correlation degree with a preset correlation degree threshold;

[0048] When the correlation degree is not lower than the preset correlation degree threshold, it is determined that there is a correlation relationship between the two windbreaker feature point cloud data corresponding to the correlation degree not lower than the preset correlation degree threshold;

[0049] Correlate the windbreaker feature point cloud data with a correlation relationship.

[0050] Preferably, setting the correlation degree threshold using each windbreaker feature point cloud data in the target point cloud data set includes:

[0051] Retrieve the cosine similarity of each windbreaker feature point cloud data in the target point cloud data set;

[0052] Obtain the median of the cosine similarity and the standard deviation of the cosine similarity from the cosine similarity of each windbreaker feature point cloud data in the target point cloud data set;

[0053] Perform a ratio process on the standard deviation of the cosine similarity and the median of the cosine similarity to obtain the coefficient of variation of the cosine similarity, and the coefficient of variation is used to reflect the degree of dispersion of the similarity distribution;

[0054] Retrieve the three-axis span corresponding to each windbreaker feature point cloud data in the target point cloud data set to obtain the average value of the three-axis span corresponding to the entire target point cloud data set;

[0055] Set the correlation degree threshold using the average value of the three-axis span in combination with the coefficient of variation of the cosine similarity;

[0056] Among them, the correlation degree threshold is obtained through the following formula:

[0057]

[0058] Among them, G y represents the correlation degree threshold; S z represents the median of the chord similarity; S k represents the coefficient of variation of the cosine similarity; K p represents the average value of the three-axis span corresponding to the entire target point cloud data set.

[0059] Preferably, the virtual 3D model of the windbreaker is precisely matched with the anthropometric data to adapt the user's body to the windbreaker, including:

[0060] Use a 3D scanning device to comprehensively scan the user's body, collect data on the user's height, weight, shoulder width, chest circumference, waist circumference, and hip circumference, and determine the user's anthropometric data;

[0061] Convert the user's anthropometric data so that the user's anthropometric data is converted into a standardized format, determine the standardized anthropometric data, and precisely match the virtual 3D model of the windbreaker with the anthropometric data;

[0062] Among them, according to the anthropometric data, adjust the dimensions of the virtual 3D model of the windbreaker, including the length of the clothes, the length of the sleeves, and the chest circumference. Among them, use geometric transformation techniques such as scaling and deformation to make the size of the windbreaker presented by the virtual 3D model of the windbreaker match the user's body;

[0063] Among them, according to the anthropometric data, adjust the shape of the virtual 3D model of the windbreaker so that the shape of the windbreaker presented by the virtual 3D model of the windbreaker matches the user's body, fits the user's body curve, and adapts the user's body to the windbreaker.

[0064] Preferably, the user conducts a virtual 3D fitting experience of the windbreaker and simulates the texture of the clothing and the fitting actions to view the dynamic effect of the windbreaker, including:

[0065] Real-time generate the virtual reality 3D fitting effect of the user wearing the windbreaker, and view the 360-degree user fitting effect by controlling the rotation;

[0066] Based on motion capture technology, the user simulates walking and running actions in the virtual environment, and views the dynamic effect of the windbreaker in real time to judge whether the windbreaker can meet the user's needs for walking and running actions;

[0067] Based on the physics engine, simulate the texture, weight, and drapability of the windbreaker, and view the display of the windbreaker in real time to judge whether the windbreaker can meet the user's needs for texture, weight, and drapability;

[0068] When the user's needs cannot be met, optimize and adjust the virtual 3D model of the windbreaker according to the user's needs to form a closed-loop feedback, so that the windbreaker presented by the virtual 3D model of the windbreaker meets the user's needs.

[0069] Preferably, it further includes:

[0070] A size automatic recommendation unit configured to automatically recommend the best virtual reality 3D fitting size of the windbreaker to the user according to the anthropometric data;

[0071] The effect display and sharing unit is configured to visually display and share the user's try-on effect;

[0072] Among them, automatically recommending the best virtual 3D try-on size of the windbreaker for the user according to the body measurement data includes:

[0073] Collecting the historical data of virtual 3D try-on of the windbreaker, dividing the collected historical data of virtual 3D try-on of the windbreaker, and determining the training set and the test set;

[0074] Based on machine learning technology, using the training set to train the machine learning model, enabling the machine learning model to autonomously learn the virtual 3D try-on size recommendation behavior of the windbreaker, and determining the virtual 3D try-on size recommendation model of the windbreaker;

[0075] Based on the test set, performing performance testing on the virtual 3D try-on size recommendation model of the windbreaker, evaluating whether the virtual 3D try-on size recommendation model of the windbreaker can achieve the expected effect; according to the evaluation results, adjusting the parameters of the virtual 3D try-on size recommendation model of the windbreaker and continuously optimizing to determine the best virtual 3D try-on size recommendation model of the windbreaker;

[0076] Inputting the body measurement data into the best virtual 3D try-on size recommendation model of the windbreaker, analyzing the body measurement data according to the best virtual 3D try-on size recommendation model of the windbreaker, and automatically recommending the best virtual 3D try-on size of the windbreaker to the user;

[0077] Among them, visually displaying and sharing the user's try-on effect includes:

[0078] Obtaining the user's try-on effect, and after obtaining the user's consent, visually displaying the user's try-on effect in real time for other users to view the try-on effect of the windbreaker, and sharing the user's try-on effect to the social platform for other users to make suggestions for the user's try-on effect.

[0079] Compared with the prior art, the beneficial effects of the present invention are:

[0080] The present invention constructs a virtual 3D model of the windbreaker by collecting and processing the real-time point cloud data of the windbreaker, allowing users to have a try-on experience in a virtual environment. By accurately matching the virtual 3D model of the windbreaker with the body measurement data, the user's body is adapted to the windbreaker, ensuring the accuracy and personalization of the try-on. According to the virtual 3D model of the windbreaker, the user can have a virtual 3D try-on experience of the windbreaker, simulate the texture of the clothing and the try-on actions, and view the dynamic effect of the windbreaker, improving the realism of the try-on. According to the body measurement data, the best virtual 3D try-on size of the windbreaker can be automatically recommended to the user, which can better meet the user's try-on needs, effectively improve the shopping experience of consumers, and reduce the return rate. Description of the Drawings

[0081] Figure 1 This is a module diagram of the virtual 3D fitting system for windbreakers of the present invention. Detailed implementation manners

[0082] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0083] In order to solve the problem that the existing system cannot enable users to have a fitting experience in a virtual environment, resulting in a poor shopping experience for consumers and a high return rate, please refer to Figure 1 , the following technical solutions are provided in this embodiment:

[0084] A virtual 3D fitting system for windbreakers includes: a 3D model construction unit, a human body model matching unit, a virtual 3D fitting unit, a size automatic recommendation unit, and an effect display and sharing unit.

[0085] Specifically, through the interactive communication between the 3D model construction unit, the human body model matching unit, the virtual 3D fitting unit, the size automatic recommendation unit, and the effect display and sharing unit, users can be allowed to have a fitting experience in a virtual environment, ensuring the accuracy and personalization of the fitting, improving the realism of the fitting, better meeting the fitting needs of users, effectively enhancing the shopping experience of consumers, and reducing the return rate.

[0086] Among them, the 3D model construction unit is used to collect the real-time point cloud data of the windbreaker and construct a virtual 3D model of the windbreaker after processing;

[0087] In this embodiment, collecting the real-time point cloud data of the windbreaker includes:

[0088] Based on 3D scanning technology, a 3D laser scanner is used to perform real-time scanning and collection of the style of the windbreaker to obtain the style point cloud data of the windbreaker;

[0089] Based on 3D scanning technology, a 3D laser scanner is used to perform real-time scanning and collection of the texture of the windbreaker to obtain the texture point cloud data of the windbreaker;

[0090] Based on 3D scanning technology, a 3D laser scanner is used to perform real-time scanning and collection of the fabric of the windbreaker to obtain the fabric point cloud data of the windbreaker;

[0091] Among them, according to the style point cloud data, texture point cloud data, and fabric point cloud data of the windbreaker, the real-time point cloud data of the windbreaker is determined.

[0092] In this embodiment, processing the real-time point cloud data of the windbreaker includes:

[0093] Cleaning the real-time point cloud data of the windbreaker to remove the noise data that is useless for the virtual 3D fitting of the windbreaker in the real-time point cloud data of the windbreaker;

[0094] Checking and identifying the real-time point cloud data of the windbreaker to determine whether there are duplicate values, missing values, and abnormal values that are useless for the virtual 3D fitting of the windbreaker in the real-time point cloud data of the windbreaker, and deleting the duplicate values, missing values, and abnormal values that are useless for the virtual 3D fitting of the windbreaker existing in the real-time point cloud data of the windbreaker;

[0095] Extracting features from the real-time point cloud data of the windbreaker, extracting the features useful for the virtual 3D fitting of the windbreaker from the real-time point cloud data of the windbreaker, and determining the feature point cloud data of the windbreaker.

[0096] In this embodiment, constructing a virtual 3D model of the windbreaker includes:

[0097] Associating the feature point cloud data of the windbreaker to determine the associated set of the feature point cloud data of the windbreaker;

[0098] Based on the virtual 3D fitting requirements of the windbreaker, according to the associated set of the feature point cloud data of the windbreaker, selecting 3D modeling software, and based on the selected 3D modeling software, performing 3D modeling on the associated set of the feature point cloud data of the windbreaker to construct a virtual 3D model of the windbreaker with a 1:1 reduction degree;

[0099] Performing detail carving and texture rendering on the virtual 3D model of the windbreaker, using carving tools to add wrinkles, stitches, and texture details, assigning fabric materials to different parts of the windbreaker, setting roughness, metallicity, and normal parameters, and rendering the texture, accurately mapping it to the model surface to determine a realistic virtual 3D model of the windbreaker.

[0100] Specifically, associating the feature point cloud data of the windbreaker to determine the associated set of the feature point cloud data of the windbreaker includes:

[0101] Scanning all the feature point cloud data of the windbreaker and extracting the feature data of each feature point cloud data of the windbreaker;

[0102] Using the feature data to obtain the cosine similarity value between every two feature point cloud data among all the feature point cloud data of the windbreaker;

[0103] Comparing the cosine similarity value between every two feature point cloud data with a preset similarity threshold;

[0104] Filter out the feature point cloud data of windbreakers with cosine similarity values less than the preset similarity threshold to form a target point cloud data set;

[0105] For each piece of feature point cloud data of the windbreaker in the target point cloud data set, obtain the projection value of each piece of feature point cloud data of the windbreaker projected onto the main axis of its corresponding oriented bounding box (OBB);

[0106] Obtain the main axis components corresponding to the three-axis data of each piece of feature point cloud data of the windbreaker from the projection values;

[0107] Use the main axis components to obtain the maximum and minimum values of each piece of feature point cloud data of the windbreaker on the main axis;

[0108] Associate the feature point cloud data of the windbreaker according to the cosine similarity value corresponding to each piece of feature point cloud data of the windbreaker and its maximum and minimum values on the main axis, and determine the associated set of feature point cloud data of the windbreaker.

[0109] The technical effects of the above technical solution are as follows: By performing feature extraction and cosine similarity calculation on the feature point cloud data of the windbreaker, the similarity degree between different point cloud data can be accurately quantified. This association method based on feature similarity avoids the limitations of traditional methods that only rely on geometric position or simple distance, and significantly improves the accuracy of association. By screening the point cloud data with cosine similarity less than the preset threshold, the point cloud with too high similarity (which may belong to the same object or irrelevant noise) can be removed, reducing the possibility of incorrect association and further improving the accuracy of association. The technical solution not only considers the feature similarity of the point cloud data, but also combines the projection values projected onto the main axis of the oriented bounding box (OBB), as well as the maximum and minimum values on the main axis. The fusion of this multi-dimensional information makes the association process more robust and can handle problems such as noise, missing or deformation in the point cloud data. The preset similarity threshold can be dynamically adjusted according to the specific application scenario to adapt to the characteristics of the point cloud data under different windbreaker styles, materials or scanning conditions, thereby enhancing the adaptability and robustness of the association algorithm. By first calculating the cosine similarity and screening out the target point cloud data set, the amount of point cloud data for subsequent processing is reduced, thus improving the overall calculation efficiency. Using the main axis of the oriented bounding box (OBB) for projection and component extraction simplifies the processing process of the point cloud data and avoids complex geometric calculations, further improving the calculation speed. The windbreaker has complex geometric shapes and texture features. This technical solution can effectively handle the point cloud data association problem in these complex scenarios by comprehensively using methods such as feature similarity, main axis projection and component extraction. Whether it is high-precision scanning or low-precision scanning, this technical solution can adapt to different scanning conditions by adjusting the similarity threshold and main axis projection strategy to ensure the accuracy and stability of the point cloud data association. Through accurate point cloud data association, detailed geometric information and texture features of the windbreaker can be obtained, providing rich data support for designers and promoting the intelligence and personalization of windbreaker design. During the manufacturing process, accurate point cloud data association can be used for quality control and inspection to ensure that the dimensional accuracy and appearance quality of the windbreaker meet the design requirements, improving the manufacturing efficiency and product quality.

[0110] Specifically, associating the windbreaker feature point cloud data according to the cosine similarity value corresponding to each windbreaker feature point cloud data and its maximum and minimum values on the main axis includes:

[0111] Extracting the cosine similarity value corresponding to each windbreaker feature point cloud data and its maximum and minimum values on the main axis;

[0112] Combining the cosine similarity value corresponding to each windbreaker feature point cloud data and its maximum and minimum values on the main axis into a dual-domain feature vector;

[0113] Among them, the dual-domain feature vector is obtained through the following formula:

[0114]

[0115] Among them, D represents the dual-domain feature vector corresponding to the point cloud data of each windbreaker feature point; S represents the cosine similarity value corresponding to the point cloud data of each windbreaker feature point; X max and X min respectively represent the maximum and minimum values on the main axis corresponding to the point cloud data of each windbreaker feature point; a represents the base of the logarithmic function, and the base is determined according to actual application requirements; specifically, X max -X min represents the geometric span of the point cloud in the directions of each axis of the OBB; log a (1 + X max -X min ) represents logarithmic compression of the geometric span to suppress the excessive influence of large-scale differences on the similarity; represents the position center offset of the point cloud in the directions of each axis of the OBB;

[0116] Construct a three-dimensional projection tensor corresponding to the point cloud data of each windbreaker feature point for each windbreaker feature point cloud data;

[0117] Among them, the three-dimensional projection tensor is obtained through the following formula:

[0118]

[0119] Among them, H represents the three-dimensional projection tensor corresponding to the point cloud data of each windbreaker feature point; K x , K y and K z respectively represent the geometric spans of the point cloud data of each windbreaker feature point on the x-axis, y-axis, and z-axis; S represents the cosine similarity value corresponding to the point cloud data of each windbreaker feature point; specifically, represents the ratio of the geometric span on the x-axis to the similarity, quantifying the spatial scale corresponding to the unit similarity; represents the density of the similarity on the geometric span of the y-axis; represents the inverse proportional coupling of the z-axis span to the square root of the similarity, enhancing the sensitivity to the scenario of small similarity and large span; S 0.5 represents non-linear amplification of low similarity values;

[0120] Generate the correlation degree between every two windbreaker feature point cloud data by using the three-dimensional projection tensor in combination with the dual-domain feature vector;

[0121] Among them, the correlation degree between every two windbreaker feature point cloud data is obtained through the following formula:

[0122]

[0123] Among them, G ij represents the correlation degree between the i-th windbreaker feature point cloud data and the j-th windbreaker feature point cloud data; D i and H i respectively represent the dual-domain feature vector and the three-dimensional projection tensor corresponding to the i-th windbreaker feature point cloud data; D j and H j respectively represent the dual-domain feature vector and the three-dimensional projection tensor corresponding to the j-th windbreaker feature point cloud data; X maxi and X mini respectively represent the maximum and minimum values on the main axis corresponding to the i-th windbreaker feature point cloud data; X maxj and X minj respectively represent the maximum and minimum values on the main axis corresponding to the j-th windbreaker feature point cloud data; Specifically, ||H i *D j ||2 refers to the L2 norm between the dual-domain feature vector and the three-dimensional projection tensor, indicating the energy intensity of the feature of point cloud j in the projection space of point cloud i; σ(H j *D i ) refers to the corresponding value after the Sigmoid function processing of the dual-domain feature vector and the three-dimensional projection tensor, which is used to compress the projection value of point cloud i in the space of point cloud j to (0, 1), indicating the correlation confidence probability; represents the geometric mean cube root of the geometric span of the two clouds, which is used to normalize the correlation intensity;

[0124] Compare the correlation degree with a preset correlation degree threshold;

[0125] When the correlation degree is not lower than the preset correlation degree threshold, it is determined that there is a correlation relationship between the two windbreaker feature point cloud data corresponding to the correlation degree not lower than the preset correlation degree threshold;

[0126] Associate the windbreaker feature point cloud data with a correlation relationship.

[0127] The technical effects of the above technical solution are as follows: By combining the cosine similarity value with the maximum and minimum values on the main axis to construct a dual-domain feature vector, the characteristics of the feature point cloud data of the windbreaker can be more comprehensively characterized, making the association judgment more accurate. The cosine similarity reflects the similarity degree between point cloud data, while the maximum and minimum values on the main axis provide important information on the spatial distribution of point cloud data. The combination of the two effectively improves the accuracy of association. Using the three-dimensional projection tensor, the point cloud data is further described from two dimensions of geometric span and similarity, enabling the association judgment to consider not only the similarity of point cloud data but also its distribution characteristics in three-dimensional space, thereby improving the accuracy of association. The combination of the dual-domain feature vector and the three-dimensional projection tensor enables the association algorithm to comprehensively consider multiple-dimensional information of point cloud data, including similarity, spatial distribution, etc., thus enhancing the robustness of the algorithm to interference factors such as noise, occlusion, and deformation. By constructing the dual-domain feature vector and the three-dimensional projection tensor, the association problem of point cloud data is transformed into an operation problem of vectors and tensors, and the use of an efficient matrix operation library can significantly improve the calculation efficiency. By setting the association degree threshold, the point cloud data with an association relationship can be quickly screened out, avoiding unnecessary calculation overhead and further improving the calculation efficiency. Since there are various styles of windbreakers and the characteristics of point cloud data are different, this technical solution can adapt to the association problem of point cloud data under different windbreaker styles by comprehensively using multiple-dimensional information such as cosine similarity, the maximum and minimum values on the main axis, and the three-dimensional projection tensor. The windbreaker may have complex geometric shapes and texture features. This technical solution can effectively handle the association problem of point cloud data in these complex scenarios by constructing the dual-domain feature vector and the three-dimensional projection tensor, ensuring the accuracy and stability of the association. Through accurate association of point cloud data, detailed geometric information and texture features of the windbreaker can be obtained, providing rich data support for designers and promoting the intelligence and personalization of windbreaker design. During the manufacturing process, accurate association of point cloud data can be used for quality control and inspection to ensure that the dimensional accuracy and appearance quality of the windbreaker meet the design requirements, improving the manufacturing efficiency and product quality. At the same time, this technical solution also helps to achieve the automation and intelligence of windbreaker manufacturing, reducing labor costs and increasing production efficiency.

[0128] Specifically, setting the association degree threshold using each piece of feature point cloud data of the windbreaker in the target point cloud data set includes:

[0129] Retrieving the cosine similarity of each piece of feature point cloud data of the windbreaker in the target point cloud data set;

[0130] Obtaining the median and standard deviation of the cosine similarity from the cosine similarity of each piece of feature point cloud data of the windbreaker in the target point cloud data set;

[0131] The ratio of the standard deviation of the cosine similarity to the median of the cosine similarity is processed to obtain the coefficient of variation of the cosine similarity, and the coefficient of variation is used to reflect the degree of dispersion of the similarity distribution;

[0132] Retrieve the average value of the three-axis spans corresponding to the overall target point cloud data set by obtaining the three-axis spans corresponding to each feature point cloud data of the windbreaker in the target point cloud data set;

[0133] Set the correlation threshold by combining the average value of the three-axis spans with the coefficient of variation of the cosine similarity;

[0134] Among them, the correlation threshold is obtained through the following formula:

[0135]

[0136] Among them, G y represents the correlation threshold; S z represents the median of the chord similarity; S k represents the coefficient of variation of the cosine similarity; K p represents the average value of the three-axis spans corresponding to the overall target point cloud data set. Specifically, the median of the chord similarity reflects the overall similarity benchmark of the data set; the coefficient of variation S b / S z represents the degree of dispersion of the similarity distribution, where S b represents the standard deviation of the cosine similarity; the average value K of the three-axis spans corresponding to the overall target point cloud data set p is used to quantify the geometric distribution characteristics of the point cloud data and incorporate it into the dynamic setting of the correlation threshold, thereby improving the accuracy and robustness of point cloud association; tanh() is used to limit the influence of the sum of the spatial spans to (-1, 1) to avoid interference from extreme values;

[0137] The technical effects of the above technical solution are as follows: By calculating the median of cosine similarity, the standard deviation of cosine similarity, and the coefficient of variation, and introducing the average value of the three-axis span of the target point cloud data set, the dynamic adjustment of the correlation threshold is achieved. This adaptive mechanism enables the threshold to vary flexibly according to the specific distribution characteristics of the point cloud data, avoiding the limitations of a fixed threshold in complex scenarios. By comprehensively utilizing the statistical features (median, standard deviation, coefficient of variation) and geometric features (average value of the three-axis span) of cosine similarity, the threshold setting takes into account both data similarity and spatial distribution characteristics, significantly enhancing the comprehensiveness and rationality of the threshold setting. The coefficient of variation effectively quantifies the dispersion degree of cosine similarity and can distinguish the similarity distribution characteristics of different data sets. Combining the median and standard deviation can more accurately describe the data distribution pattern, thereby improving the accuracy of correlation judgment. The average value of the three-axis span reflects the spatial distribution range of the point cloud data. By incorporating it into the threshold calculation, the sensitivity of correlation judgment to geometric characteristics is enhanced, which helps to identify point cloud data with significantly different spatial distributions. The threshold setting based on statistical features has a natural resistance to noise and outliers, can reduce the influence of extreme values on correlation judgment, and ensure the stability of the algorithm in complex environments. The adaptive threshold mechanism enables the algorithm to adapt to the point cloud data characteristics under different styles, materials, and scanning conditions of windbreakers, avoiding the subjectivity and limitations of manually setting thresholds. By presetting the dynamic threshold, point cloud data with a correlation relationship can be quickly screened out, reducing unnecessary computational overhead. The dynamic threshold setting enables the algorithm to adjust the calculation strategy according to the data characteristics, improving the utilization efficiency of computing resources. Combining geometric features (three-axis span) and similarity distribution characteristics enables the algorithm to handle the correlation problems of point cloud data with complex geometric shapes and texture features. Through the adaptive threshold mechanism, the algorithm can adapt to point cloud data under different scanning precisions, ensuring the accuracy and stability of the correlation. Accurate point cloud data correlation provides rich data support for windbreaker design, helping to achieve automatic adjustment and optimization of design parameters. During the manufacturing process, the dynamic correlation threshold can be used for quality control and inspection to ensure that the dimensional accuracy and appearance quality of the windbreaker meet the design requirements, improving manufacturing efficiency and product quality. Clearly showing the threshold calculation process through formulas enhances the interpretability of the algorithm, facilitating subsequent maintenance and optimization. The threshold setting module can be independent of other algorithm parts, facilitating separate adjustment and testing, and improving the maintainability of the algorithm. The threshold setting method based on statistical features and geometric features has a low computational complexity and is suitable for the rapid processing of large-scale point cloud data. The dynamic threshold mechanism enables the algorithm to adapt to the point cloud data characteristics under different application scenarios without the need for customized adjustment for specific scenarios. By continuously accumulating the point cloud data characteristics, the threshold calculation formula can be further optimized to improve the generalization performance of the algorithm. Accurate point cloud data correlation provides a high-precision data foundation for windbreaker design, supporting intelligent applications such as parametric design and virtual try-on.The dynamic correlation threshold can be used in an automated quality inspection system to achieve intelligent control of the manufacturing process of windbreakers.

[0138] It should be noted that a high-precision virtual 3D model of the windbreaker is created through 3D modeling software to ensure the authenticity of details and support the rendering of different materials and colors. Users can adjust the appearance of the windbreaker and are allowed to have a try-on experience in a virtual environment.

[0139] Among them, the human body model matching unit is used to accurately match the virtual 3D model of the windbreaker with the human body measurement data, so that the user's body fits the windbreaker.

[0140] In this embodiment, accurately matching the virtual 3D model of the windbreaker with the human body measurement data so that the user's body fits the windbreaker includes:

[0141] Using a 3D scanning device to comprehensively scan the user's body, collecting the user's height, weight, shoulder width, chest circumference, waist circumference and hip circumference data, and determining the user's human body measurement data.

[0142] Converting the user's human body measurement data to a standardized format to determine the standardized human body measurement data, and accurately matching the virtual 3D model of the windbreaker with the human body measurement data.

[0143] Among them, according to the human body measurement data, adjusting the size of the virtual 3D model of the windbreaker, including the length of the clothes, the length of the sleeves and the chest circumference. Among them, using geometric transformation techniques such as scaling and deformation to make the size of the windbreaker presented by the virtual 3D model of the windbreaker match the user's body.

[0144] Among them, according to the human body measurement data, adjusting the shape of the virtual 3D model of the windbreaker so that the shape of the windbreaker presented by the virtual 3D model of the windbreaker matches the user's body, making it fit the user's body curve, so that the user's body fits the windbreaker, which can ensure the accuracy and personalization of the try-on.

[0145] Among them, the virtual 3D try-on unit is used for users to have a virtual reality 3D try-on experience of the windbreaker, simulate the texture of the clothes and the try-on actions, and view the dynamic effect of the windbreaker.

[0146] In this embodiment, users have a virtual 3D try-on experience of the windbreaker, simulate the texture of the clothes and the try-on actions, and view the dynamic effect of the windbreaker, including:

[0147] Real-time generating the virtual reality 3D try-on effect of the user wearing the windbreaker, and viewing the 360-degree user try-on effect by controlling the rotation. Users can view the try-on effect from different angles.

[0148] Based on motion capture technology, users can simulate walking and running actions in a virtual environment, view the dynamic effects of the windbreaker in real time, and determine whether the windbreaker can meet the users' requirements for walking and running actions;

[0149] Based on a physics engine, simulate the texture, weight, and drapability of the windbreaker, and view the display of the windbreaker in real time to determine whether the windbreaker can meet the users' requirements for texture, weight, and drapability;

[0150] When the users' requirements cannot be met, optimize and adjust the virtual 3D model of the windbreaker according to the users' requirements to form a closed-loop feedback, so that the windbreaker presented by the virtual 3D model of the windbreaker meets the users' requirements and improves the realism of the virtual fitting.

[0151] Among them, the size automatic recommendation unit is used to automatically recommend the best virtual reality 3D fitting size of the windbreaker to the user according to the body measurement data;

[0152] In this embodiment, automatically recommending the best virtual 3D fitting size of the windbreaker to the user according to the body measurement data includes:

[0153] Collect the historical data of virtual 3D fitting of the windbreaker, and divide the collected historical data of virtual 3D fitting of the windbreaker to determine the training set and the test set;

[0154] Based on machine learning technology, use the training set to train the machine learning model, so that the machine learning model autonomously learns the behavior of recommending the virtual 3D fitting size of the windbreaker, and determines the virtual 3D fitting size recommendation model of the windbreaker;

[0155] Based on the test set, conduct a performance test on the virtual 3D fitting size recommendation model of the windbreaker to evaluate whether the virtual 3D fitting size recommendation model of the windbreaker can achieve the expected effect; according to the evaluation results, adjust the parameters of the virtual 3D fitting size recommendation model of the windbreaker, and through continuous optimization, determine the best virtual 3D fitting size recommendation model of the windbreaker;

[0156] Input the body measurement data into the best virtual 3D fitting size recommendation model of the windbreaker, analyze the body measurement data according to the best virtual 3D fitting size recommendation model of the windbreaker, and automatically recommend the best virtual 3D fitting size of the windbreaker to the user.

[0157] Among them, the effect display and sharing unit is used to visually display and share the user's fitting effect.

[0158] In this embodiment, visually displaying and sharing the user's fitting effect includes:

[0159] Obtain the user's try-on effect. After obtaining the user's consent, the user's try-on effect is displayed in a visual form in real time for other users to view the try-on effect of the windbreaker, and the user's try-on effect is shared on the social platform for other users to make suggestions on the user's try-on effect.

[0160] It should be noted that in this article, relational terms such as first and second are only used 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 term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0161] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A virtual 3D fitting system for windbreaker, characterized in that Including: A three-dimensional model construction unit configured to collect real-time point cloud data of a windbreaker, process it, determine the characteristic point cloud data of the windbreaker, correlate the characteristic point cloud data of the windbreaker according to the cosine similarity value corresponding to each characteristic point cloud data of the windbreaker and its maximum and minimum values on the main axis, determine the associated set of the characteristic point cloud data of the windbreaker, and construct a virtual three-dimensional model of the windbreaker; A human body model matching unit configured to accurately match the virtual three-dimensional model of the windbreaker with the human body measurement data to make the user's body fit the windbreaker; A virtual three-dimensional fitting unit configured to enable the user to experience virtual reality three-dimensional fitting of the windbreaker, simulate the texture of the clothing and the fitting actions, and view the dynamic effect of the windbreaker.

2. The virtual three-dimensional fitting system for a windbreaker according to claim 1, characterized in that Collecting real-time point cloud data of the windbreaker, including: Based on three-dimensional scanning technology, using a three-dimensional laser scanner to perform real-time scanning and collection of the style of the windbreaker to obtain the style point cloud data of the windbreaker; Based on three-dimensional scanning technology, using a three-dimensional laser scanner to perform real-time scanning and collection of the texture of the windbreaker to obtain the texture point cloud data of the windbreaker; Based on three-dimensional scanning technology, using a three-dimensional laser scanner to perform real-time scanning and collection of the fabric of the windbreaker to obtain the fabric point cloud data of the windbreaker; Among them, according to the style point cloud data, texture point cloud data and fabric point cloud data of the windbreaker, the real-time point cloud data of the windbreaker is determined.

3. The virtual three-dimensional fitting system for a windbreaker according to claim 1, wherein Processing the real-time point cloud data of the windbreaker, including: Cleaning the real-time point cloud data of the windbreaker to remove the noise data useless for the virtual three-dimensional fitting of the windbreaker in the real-time point cloud data of the windbreaker; Checking and identifying the real-time point cloud data of the windbreaker, judging whether there are duplicate values, missing values and abnormal values useless for the virtual three-dimensional fitting of the windbreaker in the real-time point cloud data of the windbreaker, and deleting the duplicate values, missing values and abnormal values useless for the virtual three-dimensional fitting of the windbreaker existing in the real-time point cloud data of the windbreaker; Extracting features from the real-time point cloud data of the windbreaker, extracting the features useful for the virtual three-dimensional fitting of the windbreaker from the real-time point cloud data of the windbreaker, and determining the characteristic point cloud data of the windbreaker.

4. The virtual three-dimensional fitting system for a windbreaker according to claim 3, characterized in that, Constructing a virtual three-dimensional model of the windbreaker, including: Correlating the characteristic point cloud data of the windbreaker to determine the associated set of the characteristic point cloud data of the windbreaker; Based on the requirements of virtual three-dimensional fitting of the windbreaker, according to the associated set of the characteristic point cloud data of the windbreaker, selecting three-dimensional modeling software, and based on the selected three-dimensional modeling software, performing three-dimensional modeling on the associated set of the characteristic point cloud data of the windbreaker to construct a virtual three-dimensional model of the windbreaker with a one-to-one restoration degree; Performing detail carving and texture rendering on the virtual three-dimensional model of the windbreaker, using carving tools to add wrinkles, stitches and texture details, assigning fabric materials to different parts of the windbreaker, setting roughness, metallicity and normal parameters, and rendering the texture, accurately mapping it to the model surface to determine the virtual three-dimensional model of the windbreaker based on virtual reality.

5. The virtual three-dimensional fitting system for a windbreaker according to claim 4, wherein, Correlating the characteristic point cloud data of the windbreaker to determine the associated set of the characteristic point cloud data of the windbreaker, including: Scanning all the characteristic point cloud data of the windbreaker and extracting the characteristic data of each characteristic point cloud data of the windbreaker; Using the feature data, for every two pieces of windbreaker feature point cloud data among all the windbreaker feature point cloud data, obtain the cosine similarity value between every two pieces of windbreaker feature point cloud data; Compare the cosine similarity value between every two pieces of windbreaker feature point cloud data with a preset similarity threshold; Filter out the windbreaker feature point cloud data with a cosine similarity value less than the preset similarity threshold to form a target point cloud data set; For each piece of windbreaker feature point cloud data in the target point cloud data set, obtain the projection value of each piece of windbreaker feature point cloud data projected onto the main axis of its corresponding directional bounding box; Obtain the main axis components corresponding to the three-axis data of each piece of windbreaker feature point cloud data from the projection values; Use the main axis components to obtain the maximum and minimum values of each piece of windbreaker feature point cloud data on the main axis; Associate the windbreaker feature point cloud data according to the cosine similarity value corresponding to each piece of windbreaker feature point cloud data and its maximum and minimum values on the main axis, and determine the associated set of windbreaker feature point cloud data.

6. The virtual three-dimensional fitting system for a windbreaker according to claim 5, wherein Associating the windbreaker feature point cloud data according to the cosine similarity value corresponding to each piece of windbreaker feature point cloud data and its maximum and minimum values on the main axis includes: Extract the cosine similarity value corresponding to each piece of windbreaker feature point cloud data and its maximum and minimum values on the main axis; Combine the cosine similarity value corresponding to each piece of windbreaker feature point cloud data and its maximum and minimum values on the main axis into a dual-domain feature vector; Construct a three-dimensional projection tensor corresponding to each piece of windbreaker feature point cloud data for each piece of windbreaker feature point cloud data; Use the three-dimensional projection tensor to generate the association degree between every two pieces of windbreaker feature point cloud data in combination with the dual-domain feature vector; Compare the association degree with a preset association degree threshold; When the association degree is not lower than the preset association degree threshold, it is determined that there is an association relationship between the two pieces of windbreaker feature point cloud data corresponding to the association degree not lower than the preset association degree threshold; Associate the windbreaker feature point cloud data with an association relationship.

7. The virtual three-dimensional fitting system for a windbreaker according to claim 5, characterized in that, Setting the association degree threshold using each piece of windbreaker feature point cloud data in the target point cloud data set includes: Retrieve the cosine similarity of each piece of windbreaker feature point cloud data in the target point cloud data set; Obtain the median cosine similarity and the standard deviation of the cosine similarity from the cosine similarity of each piece of windbreaker feature point cloud data in the target point cloud data set; Perform a ratio process on the standard deviation of the cosine similarity and the median cosine similarity to obtain the coefficient of variation of the cosine similarity, and the coefficient of variation is used to reflect the degree of dispersion of the similarity distribution; Retrieve the three-axis span corresponding to each piece of windbreaker feature point cloud data in the target point cloud data set to obtain the average value of the three-axis span corresponding to the entire target point cloud data set; Set the association degree threshold using the average value of the three-axis span in combination with the coefficient of variation of the cosine similarity.

8. The virtual three-dimensional fitting system for a windbreaker according to claim 1, characterized in that, Precisely match the virtual 3D model of the windbreaker with the anthropometric data to make the user's body fit the windbreaker, including: Use a 3D scanning device to comprehensively scan the user's body, collect data on the user's height, weight, shoulder width, chest circumference, waist circumference, and hip circumference, and determine the user's anthropometric data; Convert the user's anthropometric data to a standardized format, determine the standardized anthropometric data, and precisely match the virtual 3D model of the windbreaker with the anthropometric data; Among them, according to the anthropometric data, adjust the dimensions of the virtual 3D model of the windbreaker, including the length of the coat, the length of the sleeves, and the chest circumference. Among them, use geometric transformation techniques such as scaling and deformation to make the size of the windbreaker presented by the virtual 3D model of the windbreaker match the user's body; Among them, according to the anthropometric data, adjust the shape of the virtual 3D model of the windbreaker so that the shape of the windbreaker presented by the virtual 3D model of the windbreaker matches the user's body, conforms to the user's body curve, and makes the user's body fit the windbreaker.

9. The virtual three-dimensional fitting system for a windbreaker according to claim 1, characterized in that, The user conducts a virtual 3D try-on experience of the windbreaker and simulates the texture of the clothing and the try-on actions to view the dynamic effects of the windbreaker, including: Real-time generate the virtual reality 3D try-on effect of the user wearing the windbreaker, and view the 360-degree user try-on effect by controlling the rotation; Based on motion capture technology, the user simulates walking and running actions in the virtual environment, and views the dynamic effects of the windbreaker in real time to determine whether the windbreaker can meet the user's needs for walking and running actions; Based on the physics engine, simulate the texture, weight, and drapability of the windbreaker, and view the display of the windbreaker in real time to determine whether the windbreaker can meet the user's needs for texture, weight, and drapability; When the user's needs cannot be met, optimize and adjust the virtual 3D model of the windbreaker according to the user's needs to form a closed-loop feedback, so that the windbreaker presented by the virtual 3D model of the windbreaker meets the user's needs.

10. A virtual 3D fitting system for windbreakers as described in claim 1, characterized in that, It also includes: A size automatic recommendation unit configured to automatically recommend the best virtual reality 3D try-on size of the windbreaker to the user according to the anthropometric data; An effect display and sharing unit configured to visually display and share the user's try-on effect; Among them, automatically recommending the best virtual 3D try-on size of the windbreaker to the user according to the anthropometric data includes: Collect the historical data of the virtual 3D try-on of the windbreaker, divide the collected historical data of the virtual 3D try-on of the windbreaker, and determine the training set and the test set; Based on machine learning technology, use the training set to train the machine learning model, enable the machine learning model to autonomously learn the behavior of recommending the virtual 3D try-on size of the windbreaker, and determine the virtual 3D try-on size recommendation model of the windbreaker; Based on the test set, conduct performance testing on the virtual 3D try-on size recommendation model of the windbreaker, evaluate whether the virtual 3D try-on size recommendation model of the windbreaker can achieve the expected effect; according to the evaluation results, adjust the parameters of the virtual 3D try-on size recommendation model of the windbreaker and continuously optimize it to determine the best virtual 3D try-on size recommendation model of the windbreaker; Input the anthropometric data into the best virtual 3D try-on size recommendation model of the windbreaker, analyze the anthropometric data according to the best virtual 3D try-on size recommendation model of the windbreaker, and automatically recommend the best virtual 3D try-on size of the windbreaker to the user; Among them, visual display and sharing of the user's try-on effect include: Obtain the user's try-on effect. After obtaining the user's consent, the user's try-on effect is displayed in real time in a visual form for other users to view the try-on effect of the windbreaker, and the user's try-on effect is shared on the social platform for other users to make suggestions on the user's try-on effect.

Images