Intelligent fitting system and virtual fitting method

Abstract

The present invention discloses an intelligent fitting system and a virtual fitting method, providing an intelligent fitting system, the intelligent fitting system comprising: a physical measurement subsystem, comprising a physical measurement sensor, for measuring the physical information of the fitting person and generating weight parameters and physical parameters, the physical parameters including body fat percentage data and muscle mass data; a stereoscopic vision acquisition subsystem, comprising a binocular depth camera, for collecting the fitting person's body image information and generating body parameters and height parameters, the body parameters including chest circumference data, waist circumference data, hip circumference data and shoulder width data; a transparent display device, comprising a control calculation subsystem and a transparent display screen electrically connected thereto, the control calculation subsystem for obtaining the body parameters and height parameters and calculating and generating a three-dimensional image of a human body model. The fitting system of the present invention has a strong sense of presence and realism, and the establishment of the human body model can be improved according to the body fat percentage data and muscle mass data.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual fitting technology, and in particular to an intelligent fitting system and a virtual fitting method. Background Art

[0002] Trying on clothes is an essential part of the apparel retail industry and a crucial way to boost sales and acquire customer resources. Traditional offline clothing stores all have fitting rooms where customers can change clothes and see how they're trying on. While traditional fitting rooms are common in clothing retail stores, they often require customers to constantly change clothes, leading to long wait times and repetitive workloads for sales staff, resulting in poor customer retention. Smart fitting mirrors combine traditional fitting mirrors with display screens, partially displaying the effects of fitting on a person's face through computer-rendered image rendering. However, these devices suffer from noticeable artifacts and lack a realistic feel. Existing virtual fitting equipment uses dual SLR cameras to take photos of the user and then overlay the fitting effects. However, this equipment is difficult to set up, expensive, and lacks realism. Existing virtual fitting methods use Flash technology to help shoppers choose the right outfit. However, the clothing and figures are presented as flat photos. Even if they can be rotated, they are simply a combination of multiple flat photos, failing to achieve a realistic effect. Existing virtual 3D fitting systems are complex and challenging. Trying on 3D clothing is the biggest challenge plaguing 3D virtual reality technology. This is because once a 3D garment is placed on a 3D body, it must deform according to constraints such as the body's contours and the garment's material, thereby determining the garment's comfort level and achieving the desired fit. However, these methods all share a common problem: the sense of presence and realism in fitting is weak, and the human body model is often inaccurate. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent fitting system and a virtual fitting method, which have a strong sense of presence and reality in fitting and the establishment of a human body model can be improved according to body fat percentage data and muscle mass data.

[0004] To achieve the above objectives, the present invention provides the following technical solutions: an intelligent fitting system, comprising:

[0005] The body measurement subsystem includes a body measurement sensor for measuring the body information of the person being fitted and generating weight parameters and body parameters, wherein the body parameters include body fat percentage data and muscle mass data;

[0006] A stereoscopic vision acquisition subsystem, including a binocular depth camera, is used to collect body image information of the person being fitted and generate body parameters and height parameters, wherein the body parameters include chest circumference data, waist circumference data, hip circumference data, and shoulder width data;

[0007] A transparent display device, comprising a control and computing subsystem and a transparent display screen electrically connected thereto, wherein the control and computing subsystem is used to obtain the body shape parameters and height parameters and calculate and generate a three-dimensional image of a human body model, and the transparent display screen comprises a human-computer interaction module and a three-dimensional image display module;

[0008] The control and calculation subsystem is configured with a model improvement strategy, which includes using the weight parameters and physique parameters to calculate and correct the three-dimensional image of the human body model to generate a standard three-dimensional image of the human body model; the control and calculation subsystem is configured with a fitting database, which stores a number of three-dimensional images of clothing of different styles and sizes.

[0009] A virtual fitting method and an intelligent fitting system are provided, wherein the intelligent fitting system comprises:

[0010] The body measurement subsystem includes a body measurement sensor for measuring the body information of the person being fitted and generating weight parameters and body parameters, wherein the body parameters include body fat percentage data and muscle mass data;

[0011] A stereoscopic vision acquisition subsystem, including a binocular depth camera, is used to collect body image information of the person being fitted and generate body parameters and height parameters, wherein the body parameters include chest circumference data, waist circumference data, hip circumference data, and shoulder width data;

[0012] A transparent display device, comprising a control and calculation subsystem and a transparent display screen electrically connected thereto, wherein the control and calculation subsystem is used to obtain the body shape parameters and height parameters and calculate and generate a three-dimensional image of a human body model;

[0013] The control and computing subsystem is configured with a model improvement strategy, wherein the model improvement strategy includes calculating and modifying a three-dimensional image of a human body model using the weight parameters and the physique parameters to generate a standard three-dimensional image of a human body model; the control and computing subsystem is configured with a fitting database, wherein the fitting database stores three-dimensional images of a plurality of garments of different styles and sizes; the three-dimensional garment images are configured with a cloth simulation algorithm and a physics engine, wherein the cloth simulation algorithm is configured to enable the three-dimensional garment images to deform within a preset range;

[0014] The virtual fitting method includes a pre-judgment step, a first virtual dressing step, and a second virtual dressing step. The pre-judgment step is configured with a pre-judgment strategy, the pre-judgment strategy includes a preset first difference and a second difference, and the pre-judgment step includes:

[0015] Step S0: completing the clothing selection process through the human-computer interaction module;

[0016] Step S1: Expanding the selected three-dimensional image of clothing according to a preset expansion path and expansion size to generate a three-dimensional clothing model, forming a humanoid space within the three-dimensional clothing model, and calculating the volume of the humanoid space to generate first volume data;

[0017] Step S2: reducing the three-dimensional image of the standard human body model of the person fitting the clothes according to a preset ratio to form an initial human body model, and calculating the volume of the initial human body model to generate second volume data;

[0018] Step S3: Compare the first volume data and the second volume data according to the judgment strategy. When the second volume data is less than or equal to the first volume data and the difference between the two is greater than the first difference, output the first prompt information; when the second volume data is less than or equal to the first volume data and the difference between the two is less than or equal to the first difference, jump to the first virtual dressing step; when the second volume data is greater than the first volume data, jump to subroutine one, and make different judgments based on the material of the clothing.

[0019] The first difference value may be the sum of a preset reduction ratio value of the three-dimensional image of the human body model and a normal clothing allowance value of the person trying on clothes.

[0020] Because different types of clothing may have slightly different sizes, the purpose of the first prompt message is to remind the person trying on clothes that the size they have chosen is too large, and to suggest that the person change to a slightly smaller size to improve the efficiency of trying on clothes.

[0021] Preferably, the first virtual dressing step is used to generate a three-dimensional dressing image, and the first virtual dressing step is configured with a human body model expansion algorithm, the human body model expansion algorithm includes a preset first threshold, and the virtual dressing step includes:

[0022] Step A1: Based on the shoulder line, the initial human body model is loaded into the three-dimensional clothing model, and a reserved area is formed between the initial human body model and the three-dimensional clothing model;

[0023] Step A2: using the shoulder line as a reference, using a human body model expansion algorithm to continuously merge background points within a first threshold around the initial human body model into the initial human body model, so as to reconstruct the initial human body model into a three-dimensional form and regenerate the three-dimensional image of the standard human body model;

[0024] Step A3, ending the first virtual dressing step and generating a three-dimensional dressing image combining the three-dimensional image of the standard human body model and the three-dimensional clothing model;

[0025] Step A4, a demonstration step, includes starting the physical engine and displaying the three-dimensional image of clothing on the three-dimensional image display module.

[0026] Preferably, the subroutine 1 includes obtaining material information of the clothing, classifying it into one of elastic material and inelastic material according to a preset classification standard, outputting a third prohibition message when the clothing material is inelastic material, and jumping to the second virtual dressing step when the clothing material is elastic material and the difference between the second volume data and the first volume data is less than or equal to the second difference; and outputting a second prompt message when the clothing material is elastic material and the difference between the second volume data and the first volume data is greater than the second difference.

[0027] The purpose of the second prompt message is to remind the person that although the garment is made of elastic material, it will be tight when worn by the person, and the person is advised to change to a slightly larger size.

[0028] The purpose of the third prohibition message is to remind the person trying on the clothes that the clothing is made of non-elastic material and cannot be worn according to the person's actual body shape. The person trying on the clothes needs to change to a larger size.

[0029] Preferably, the first virtual dressing step is used to generate a three-dimensional dressing image, and the first virtual dressing step is configured with a human body model expansion algorithm, the human body model expansion algorithm includes a preset first threshold, and the virtual dressing step includes:

[0030] Step B1: the elastic material includes a preset stretching threshold, the three-dimensional clothing model is expanded according to the stretching threshold, and the volume of the expanded humanoid space is calculated to update the first volume data;

[0031] Step B2: Based on the shoulder line, the initial human body model is loaded into the expanded three-dimensional clothing model, and a reserved area is formed between the initial human body model and the expanded three-dimensional clothing model;

[0032] Step B3: using the shoulder line as a reference, using a human body model expansion algorithm to continuously merge background points within a first threshold around the initial human body model into the initial human body model, so as to reconstruct the initial human body model into a three-dimensional form and regenerate the three-dimensional image of the standard human body model;

[0033] Step B4, ending the first virtual dressing step and generating a dressing three-dimensional image that combines the three-dimensional image of the standard human body model with the expanded three-dimensional clothing model;

[0034] Step B5, a demonstration step, includes starting the physical engine and displaying the three-dimensional image of clothing on the three-dimensional image display module.

[0035] Preferably, the first prompt message and the second prompt message include a confirmation operation option and a cancel operation option, the first virtual dressing step responds to the confirmation operation option of the first prompt message, and the second virtual dressing step responds to the confirmation operation option of the second prompt message; when receiving the cancel operation option of the first prompt message or the cancel operation option of the second prompt message or the third prohibition message, jump to step S0.

[0036] Preferably, the 3D image display module includes a multi-clothing comparison submodule for simultaneously comparing 3D images of multiple clothing items. When trying on clothes, a user may select multiple items and then choose the one or more items they like best. The multi-clothing comparison submodule can intuitively compare the 3D images of multiple items simultaneously, facilitating selection for the user.

[0037] Preferably, the demonstration step in step A4 or step B4 is configured with a motion tracking strategy, the motion tracking strategy includes establishing a three-dimensional human skeleton dynamic model, and the three-dimensional human skeleton dynamic model is configured with joint constraints and physiological constraints;

[0038] The motion tracking strategy is configured with a motion database, and the motion database stores a plurality of motion training samples;

[0039] The motion tracking strategy includes a motion tracking method, which includes acquiring the fitter's motion in real time through a stereoscopic vision acquisition subsystem and performing correlation calculations on it with multiple motion training samples in the motion database, and selecting the motion training sample with the highest correlation to display it on a three-dimensional image display module.

[0040] Preferably, the human-computer interaction module includes a walking vibration simulation module, a wind field simulation module and a clothing selection module. The clothing selection module includes clothing classifications of different materials and types. The walking vibration simulation module is used to simulate the vibration of the chest when walking when the selected clothing type is women's clothing. The walking vibration simulation module is configured with a vibration simulation strategy, and the vibration simulation strategy includes generating vibration amplitudes of different sizes according to the weight, chest circumference and stride of the person trying on clothes. The wind field simulation module is used to generate wind fields of different directions and different wind speeds.

[0041] Preferably, the three-dimensional image display module of the transparent display screen is arranged on one side or both sides of the transparent display screen, and the transparent display screen includes but is not limited to a transparent LED screen, an LCD screen, an OLED screen, an AMOLED screen or a self-luminous transparent screen.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] By setting a physique measurement subsystem and a model improvement strategy, the present invention can correct the three-dimensional image of the human body model, so that the three-dimensional image of the human body model is more consistent with the actual body shape of the person trying on clothes, thereby improving the accuracy of human body modeling.

[0044] The fitting method of the present invention reduces the three-dimensional image of a human body model according to a preset ratio to form an initial human body model, loads the initial human body model into a three-dimensional clothing model that has been expanded, and then uses a human body model expansion algorithm to continuously reconstruct the initial human body model to regenerate a three-dimensional image of a standard human body model, thereby simulating the actual fitting process and establishing a three-dimensional dressing image that combines the three-dimensional image of the standard human body model with the three-dimensional clothing model. This fitting method has high computational processing efficiency and is not prone to errors.

[0045] The three-dimensional image of clothing of the present invention is configured with a cloth simulation algorithm and a physical engine. The cloth simulation algorithm is configured to enable the three-dimensional image of clothing to deform within a preset range, and the physical engine can simulate the effect of gravity. The combination of the cloth simulation algorithm and the physical engine can show how clothing made of different fabrics fits the human body model under the action of gravity.

[0046] The fitting method of the present invention is equipped with a pre-judgment step. By pre-comparing the volume of the human-shaped space formed inside the three-dimensional clothing model with the volume of the initial human body model, it is preliminarily judged whether the size selected by the user is appropriate, and a corresponding prompt message or prohibition message is sent to the human-computer interaction module for display, thereby reducing unnecessary calculations of the fitting system and saving customers' fitting time.

[0047] The fitting method of the present invention classifies clothing into one of elastic materials or inelastic materials according to preset classification standards, and formulates corresponding virtual dressing steps respectively, which can better meet the different fitting needs of customers.

[0048] The human-computer interaction module of the present invention includes a walking vibration simulation module, a wind field simulation module, and a clothing selection module. The clothing selection module allows the selection of clothing of different materials and types. The walking vibration simulation module is used to simulate the vibration of the chest during walking when the selected clothing type is women's clothing. In particular, when the clothing is a strapless dress, female users are more concerned about whether the clothing will be exposed during actual walking. The walking vibration simulation module is equipped with a vibration simulation strategy, which includes generating different vibration amplitudes according to the weight, chest circumference, and stride length of the user. The wind field simulation module can be used to generate wind fields of different directions and wind speeds. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a schematic diagram of a circuit connection block diagram of a first embodiment of an intelligent fitting system of the present invention;

[0050] Figure 2 This is a schematic diagram of a circuit connection block diagram of a second embodiment of an intelligent fitting system of the present invention;

[0051] Figure 3 Schematic diagram of the flow of the pre-judgment step in the virtual fitting method of the present invention;

[0052] Figure 4 Schematic diagram of the first virtual dressing step in the virtual fitting method of the present invention;

[0053] Figure 5 2 is a flow chart of the second virtual dressing step in the virtual fitting method of the present invention. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] like Figure 1 As shown, the first embodiment provided by the present invention is an intelligent fitting system, comprising: a physical measurement subsystem, including a physical measurement sensor, for measuring the physical information of the person fitting the clothes and generating weight parameters and physical parameters, wherein the physical parameters include body fat percentage data and muscle mass data;

[0056] Based on the body shape analysis method combined with physical sensors (such as weight, body fat sensors, etc.), the accuracy of the human body 3D model is further improved by analyzing and processing the data collected by the physical sensors.

[0057] A stereoscopic vision acquisition subsystem, including a binocular depth camera, is used to collect body image information of the person being fitted and generate body parameters and height parameters, wherein the body parameters include chest circumference data, waist circumference data, hip circumference data, and shoulder width data;

[0058] A transparent display device, comprising a control and computing subsystem and a transparent display screen electrically connected thereto, wherein the control and computing subsystem is used to obtain the body shape parameters and height parameters and calculate and generate a three-dimensional image of a human body model, and the transparent display screen comprises a human-computer interaction module and a three-dimensional image display module;

[0059] The human-computer interaction mode includes but is not limited to touch mode, voice mode, gesture control and the like.

[0060] The three-dimensional image display module of the transparent display screen can be set on one side or both sides of the transparent display screen. The transparent display screen used is a single-sided or double-sided transparent display screen. The person trying on clothes can directly see the effect of his or her fitting on this side. If a double-sided transparent display screen is used, the reference person can also see the fitting effect on the other display side and put forward reference opinions, thereby increasing the fun of fitting on clothes. The transparent display screen includes but is not limited to a transparent LED screen, an LCD screen, an OLED screen, an AMOLED screen or a self-luminous transparent screen.

[0061] The control and calculation subsystem is configured with a model improvement strategy, which includes using the weight parameters and physique parameters to calculate and correct the three-dimensional image of the human body model to generate a standard three-dimensional image of the human body model; the control and calculation subsystem is configured with a fitting database, which stores a number of three-dimensional images of clothing of different styles and sizes.

[0062] like Figure 2 As shown, a circuit connection block diagram of a second embodiment of an intelligent fitting system; the human-computer interaction module includes a walking vibration simulation module, a wind field simulation module and a clothing selection module, the clothing selection module includes clothing classifications of different materials and types, the walking vibration simulation module is used to simulate the vibration of the chest when walking when the selected clothing type is female clothing, the walking vibration simulation module is configured with a vibration simulation strategy, and the vibration simulation strategy includes generating vibration amplitudes of different sizes according to the weight, chest circumference and stride of the person fitting the clothes, and the wind field simulation module is used to generate wind fields of different directions and different wind speeds.

[0063] The clothing selection module allows users to select clothing of different materials and different types. The walking vibration simulation module is used to simulate the vibration of the chest when walking when the selected clothing type is female clothing. Especially when the clothing is a strapless dress, female users are more concerned about whether the clothing will be exposed during actual walking. The walking vibration simulation module is equipped with a vibration simulation strategy, which includes generating vibration amplitudes of different sizes according to the weight, chest circumference and stride of the user. The stride of the user can be input through the human-computer interaction module. The wind field simulation module can be used to generate wind fields of different directions and wind speeds to enhance the sense of presence and reality of the fitting process. The wind field simulation can better fit the actual use of the clothing, making it easier for the user to observe some conditions that may occur during normal use of the clothing and make judgments.

[0064] like Figure 3FIG2 is a flow chart of the pre-judgment step in the virtual fitting method of the present invention; a virtual fitting method provides an intelligent fitting system, the intelligent fitting system comprising: a physical measurement subsystem, including a physical measurement sensor, for measuring the physical information of the fitting subject and generating weight parameters and physical parameters, the physical parameters including body fat percentage data and muscle mass data;

[0065] A stereoscopic vision acquisition subsystem, including a binocular depth camera, is used to collect body image information of the person being fitted and generate body parameters and height parameters, wherein the body parameters include chest circumference data, waist circumference data, hip circumference data, and shoulder width data;

[0066] A transparent display device, comprising a control and calculation subsystem and a transparent display screen electrically connected thereto, wherein the control and calculation subsystem is used to obtain the body shape parameters and height parameters and calculate and generate a three-dimensional image of a human body model;

[0067] The control and calculation subsystem is configured with a model improvement strategy, which includes using the weight parameters and physical parameters to calculate and correct the three-dimensional image of the human body model to generate a standard three-dimensional image of the human body model; the control and calculation subsystem is configured with a fitting database, which stores a number of three-dimensional images of clothing of different styles and sizes; the three-dimensional clothing image is configured with a fabric simulation algorithm and a physics engine, and the fabric simulation algorithm is configured to enable the three-dimensional clothing image to be deformed within a preset range; the physics engine can simulate the effect of gravity, and the combination of the fabric simulation algorithm and the physics engine can show how clothing of different fabrics fits the human body model under the action of gravity.

[0068] The virtual fitting method includes a pre-judgment step, a first virtual dressing step, and a second virtual dressing step. The pre-judgment step is configured with a pre-judgment strategy, the pre-judgment strategy includes a preset first difference and a second difference, and the pre-judgment step includes:

[0069] Step S0: completing the clothing selection process through the human-computer interaction module;

[0070] Step S1: Expanding the selected three-dimensional image of clothing according to a preset expansion path and expansion size to generate a three-dimensional clothing model, forming a humanoid space within the three-dimensional clothing model, and calculating the volume of the humanoid space to generate first volume data;

[0071] Step S2: reducing the three-dimensional image of the standard human body model of the person fitting the clothes according to a preset ratio to form an initial human body model, and calculating the volume of the initial human body model to generate second volume data;

[0072] The three-dimensional image of the human body model is reduced to a preset ratio to form an initial human body model, which is loaded into an expanded three-dimensional clothing model. Then, the initial human body model is continuously reconstructed using a human body model expansion algorithm to regenerate a three-dimensional image of the standard human body model, thereby simulating the actual fitting process and establishing a three-dimensional dressing image that combines the three-dimensional image of the standard human body model with the three-dimensional clothing model. This fitting method has high computational efficiency and is not prone to errors.

[0073] Step S3: Compare the first volume data and the second volume data according to the judgment strategy. When the second volume data is less than or equal to the first volume data and the difference between the two is greater than the first difference, output a first prompt message; when the second volume data is less than or equal to the first volume data and the difference between the two is less than or equal to the first difference, jump to the first virtual dressing step; when the second volume data is greater than the first volume data, jump to subroutine one, and make different judgments based on the material of the clothing. The pre-judgment step of the present invention preliminarily determines whether the size selected by the user is appropriate by comparing the volume of the human-shaped space formed within the three-dimensional clothing model with the volume of the initial human body model. The corresponding prompt message or prohibition message is then sent to the human-computer interaction module for display, reducing unnecessary computational complexity of the fitting system while also saving customers time in fitting.

[0074] Figure 4 This is a flow chart of the first virtual dressing step in the virtual fitting method of the present invention; the first virtual dressing step is used to generate a three-dimensional dressing image. The first virtual dressing step is configured with a human body model expansion algorithm, the human body model expansion algorithm includes a preset first threshold, and the virtual dressing step includes:

[0075] Step A1: Based on the shoulder line, the initial human body model is loaded into the three-dimensional clothing model, and a reserved area is formed between the initial human body model and the three-dimensional clothing model;

[0076] Step A2: using the shoulder line as a reference, using a human body model expansion algorithm to continuously merge background points within a first threshold around the initial human body model into the initial human body model, so as to reconstruct the initial human body model into a three-dimensional form and regenerate the three-dimensional image of the standard human body model;

[0077] Step A3, ending the first virtual dressing step and generating a three-dimensional dressing image combining the three-dimensional image of the standard human body model and the three-dimensional clothing model;

[0078] Step A4, a demonstration step, includes starting the physical engine and displaying the three-dimensional image of clothing on the three-dimensional image display module.

[0079] Figure 5The figure is a flow chart of the second virtual dressing step in the virtual fitting method of the present invention. Subroutine 1 includes obtaining material information of the garment and classifying it as either elastic or inelastic according to a preset classification standard. When the garment material is inelastic, a third prohibition message is output. When the garment material is elastic and the difference between the second volume data and the first volume data is less than or equal to the second difference, the process jumps to the second virtual fitting step. When the garment material is elastic and the difference between the second volume data and the first volume data is greater than the second difference, a second prompt message is output. Classifying the garment as either elastic or inelastic according to the preset classification standard and formulating corresponding virtual fitting steps can better meet the diverse fitting needs of customers.

[0080] Preferably, the first virtual dressing step is used to generate a three-dimensional dressing image, and the first virtual dressing step is configured with a human body model expansion algorithm, the human body model expansion algorithm includes a preset first threshold, and the virtual dressing step includes:

[0081] Step B1: the elastic material includes a preset stretching threshold, the three-dimensional clothing model is expanded according to the stretching threshold, and the volume of the expanded humanoid space is calculated to update the first volume data;

[0082] Step B2: Based on the shoulder line, the initial human body model is loaded into the expanded three-dimensional clothing model, and a reserved area is formed between the initial human body model and the expanded three-dimensional clothing model;

[0083] Step B3: using the shoulder line as a reference, using a human body model expansion algorithm to continuously merge background points within a first threshold around the initial human body model into the initial human body model, so as to reconstruct the initial human body model into a three-dimensional form and regenerate the three-dimensional image of the standard human body model;

[0084] Step B4, ending the first virtual dressing step and generating a dressing three-dimensional image that combines the three-dimensional image of the standard human body model with the expanded three-dimensional clothing model;

[0085] Step B5, a demonstration step, includes starting the physical engine and displaying the three-dimensional image of clothing on the three-dimensional image display module.

[0086] Preferably, the first prompt message and the second prompt message include a confirmation operation option and a cancel operation option, the first virtual dressing step responds to the confirmation operation option of the first prompt message, and the second virtual dressing step responds to the confirmation operation option of the second prompt message; when receiving the cancel operation option of the first prompt message or the cancel operation option of the second prompt message or the third prohibition message, jump to step S0.

[0087] Preferably, the three-dimensional image display module includes a multi-clothing comparison submodule for comparing three-dimensional images of multiple pieces of clothing simultaneously.

[0088] Preferably, the demonstration step in step A4 or step B4 is configured with a motion tracking strategy, the motion tracking strategy includes establishing a three-dimensional human skeleton dynamic model, and the three-dimensional human skeleton dynamic model is configured with joint constraints and physiological constraints;

[0089] The motion tracking strategy is configured with a motion database, and the motion database stores a plurality of motion training samples;

[0090] The motion tracking strategy includes a motion tracking method, which includes obtaining the movements of the person trying on clothes in real time through a stereoscopic vision acquisition subsystem and performing correlation calculations on them with multiple motion training samples in the motion database, selecting the motion training sample with the highest correlation and displaying it on a three-dimensional image display module, which can enhance the authenticity of the fitting process. The application of motion tracking in other fields by establishing a three-dimensional human skeleton dynamic model is relatively mature, and this part is all existing technology.

[0091] Preferably, the three-dimensional image display module of the transparent display screen is arranged on one side or both sides of the transparent display screen, and the transparent display screen includes but is not limited to a transparent LED screen, an LCD screen, an OLED screen, an AMOLED screen or a self-luminous transparent screen.

[0092] Working principle: The present invention preliminarily determines whether the size selected by the user is appropriate by comparing the volume of the human-shaped space formed inside the three-dimensional clothing model with the volume of the initial human body model in advance, and sends corresponding prompt information or prohibition information to the human-computer interaction module for display, thereby reducing the unnecessary amount of calculation of the fitting system and saving the customer's fitting time. By setting a physical measurement subsystem and a model improvement strategy, the three-dimensional image of the human body model can be corrected, so that the three-dimensional image of the human body model is more in line with the actual body shape of the person fitting the clothes, thereby improving the accuracy of the human body model modeling. In addition, the fitting method of the present invention forms an initial human body model by reducing the three-dimensional image of the human body model according to a preset ratio, and loads it into the expanded three-dimensional clothing model, and then uses the human body model expansion algorithm to continuously reconstruct it from the initial human body model to regenerate the standard human body model three-dimensional image, so as to simulate the actual fitting process and establish a dressing three-dimensional image combining the standard human body model three-dimensional image and the three-dimensional clothing model; this fitting method has high computational efficiency and is not prone to errors. The 3D image is configured with a fabric simulation algorithm and a physics engine. The fabric simulation algorithm is configured to deform the 3D garment image within a preset range, while the physics engine simulates gravity. The combination of the fabric simulation algorithm and physics engine demonstrates how garments made of different fabrics fit the human model under gravity. Garments are categorized as either elastic or inelastic according to preset classification criteria, and corresponding virtual dressing steps are developed for each, better meeting the diverse fitting needs of customers. The human-computer interaction module includes a walking vibration simulation module, a wind field simulation module, and a clothing selection module. The walking vibration simulation module simulates the vibration of the chest during walking when the selected garment type is women's clothing. This is particularly important for women trying on strapless dresses, as women are more concerned about whether the garment will be exposed during actual walking. The walking vibration simulation module is equipped with a vibration simulation strategy that generates different vibration amplitudes based on the wearer's weight, chest circumference, and stride length. The wind field simulation module generates wind fields of varying directions and speeds.

[0093] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A virtual fitting method, characterized in that: Provided is an intelligent fitting system, comprising: The body measurement subsystem includes a body measurement sensor for measuring the body information of the person being fitted and generating weight parameters and body parameters, wherein the body parameters include body fat percentage data and muscle mass data; A stereoscopic vision acquisition subsystem, including a binocular depth camera, is used to collect body image information of the person being fitted and generate body parameters and height parameters, wherein the body parameters include chest circumference data, waist circumference data, hip circumference data, and shoulder width data; A transparent display device, comprising a control and calculation subsystem and a transparent display screen electrically connected thereto, wherein the control and calculation subsystem is used to obtain the body shape parameters and height parameters and calculate and generate a three-dimensional image of a human body model; The control and computing subsystem is configured with a model improvement strategy, which includes calculating and correcting a three-dimensional image of a human body model using the weight parameters and the physique parameters to generate a standard three-dimensional image of the human body model; the control and computing subsystem is configured with a fitting database, which stores a plurality of three-dimensional images of clothing of different styles and sizes; the three-dimensional clothing image is configured with a cloth simulation algorithm and a physics engine, the physics engine is used to create a physical scene with specific parameters, and the cloth simulation algorithm is configured to enable the three-dimensional clothing image to deform within a preset range under the influence of the physical scene; The virtual fitting method includes a pre-judgment step, a first virtual dressing step, and a second virtual dressing step. The pre-judgment step is configured with a pre-judgment strategy, the pre-judgment strategy includes a preset first difference and a second difference, and the pre-judgment step includes: Step S0: The person trying on clothes selects the corresponding clothes through the human-computer interaction module, and the human-computer interaction module retrieves the corresponding three-dimensional image of the clothes from the fitting database and displays it on the three-dimensional image display module; Step S1: Expanding the selected three-dimensional image of clothing according to a preset expansion path and expansion size to generate a three-dimensional clothing model, forming a humanoid space within the three-dimensional clothing model, and calculating the volume of the humanoid space to generate first volume data; Step S2: reducing the three-dimensional image of the standard human body model of the person fitting the clothes according to a preset ratio to form an initial human body model, and calculating the volume of the initial human body model to generate second volume data; Step S3: Compare the first volume data and the second volume data according to the judgment strategy. When the second volume data is less than or equal to the first volume data and the difference between the two is greater than the first difference, output a first prompt message; when the second volume data is less than or equal to the first volume data and the difference between the two is less than or equal to the first difference, jump to the first virtual dressing step; when the second volume data is greater than the first volume data, jump to subroutine 1, and make different judgments based on the material of the clothing; the first virtual dressing step is used to generate a three-dimensional dressing image, and the first virtual dressing step is configured with a human body model expansion algorithm, and the human body model expansion algorithm includes a preset first threshold. The virtual dressing step includes: Step A1: Based on the shoulder line, the initial human body model is loaded into the three-dimensional clothing model, and a reserved area is formed between the initial human body model and the three-dimensional clothing model; Step A2: using the shoulder line as a reference, using a human body model expansion algorithm to continuously merge background points within a first threshold around the initial human body model into the initial human body model, so as to reconstruct the initial human body model into a three-dimensional form and regenerate the three-dimensional image of the standard human body model; Step A3, ending the first virtual dressing step and generating a three-dimensional dressing image combining the three-dimensional image of the standard human body model and the three-dimensional clothing model; Step A4, a demonstration step, includes starting the physical engine and displaying the three-dimensional image of clothing on the three-dimensional image display module.

2. The virtual fitting method according to claim 1, wherein: The subroutine 1 includes obtaining material information of the clothing and classifying it into one of elastic material and inelastic material according to a preset classification standard. When the clothing material is inelastic material, a third prohibition message is output; when the clothing material is elastic material and the difference between the second volume data and the first volume data is less than or equal to the second difference, the process jumps to the second virtual dressing step; when the clothing material is elastic material and the difference between the second volume data and the first volume data is greater than the second difference, a second prompt message is output.

3. The virtual fitting method according to claim 2, wherein: The first virtual dressing step is used to generate a three-dimensional dressing image. The first virtual dressing step is configured with a human body model expansion algorithm. The human body model expansion algorithm includes a preset first threshold. The virtual dressing step includes: Step B1: the elastic material includes a preset stretching threshold, the three-dimensional clothing model is expanded according to the stretching threshold, and the volume of the expanded humanoid space is calculated to update the first volume data; Step B2: Based on the shoulder line, the initial human body model is loaded into the expanded three-dimensional clothing model, and a reserved area is formed between the initial human body model and the expanded three-dimensional clothing model; Step B3: using the shoulder line as a reference, using a human body model expansion algorithm to continuously merge background points within a first threshold around the initial human body model into the initial human body model, so as to reconstruct the initial human body model into a three-dimensional form and regenerate the three-dimensional image of the standard human body model; Step B4, ending the first virtual dressing step and generating a dressing three-dimensional image that combines the three-dimensional image of the standard human body model with the expanded three-dimensional clothing model; Step B5, a demonstration step, includes starting the physical engine and displaying the three-dimensional image of clothing on the three-dimensional image display module.

4. The virtual fitting method according to claim 2, wherein: The first prompt information and the second prompt information include a confirmation operation option and a cancellation operation option, the first virtual dressing step responds to the confirmation operation option of the first prompt information, and the second virtual dressing step responds to the confirmation operation option of the second prompt information; When receiving the cancel operation option of the first prompt information or the cancel operation option of the second prompt information or the third prohibition information, the process jumps to step S0.

5. The virtual fitting method according to any one of claims 1 to 4, characterized in that: The three-dimensional image display module includes a multi-clothing comparison submodule for comparing three-dimensional images of multiple pieces of clothing at the same time.

6. The virtual fitting method according to any one of claims 1 to 3, characterized in that: The demonstration step in step A4 or step B4 is configured with a motion tracking strategy, wherein the motion tracking strategy includes establishing a three-dimensional human skeleton dynamic model, wherein the three-dimensional human skeleton dynamic model is configured with joint constraints and physiological constraints; The motion tracking strategy is configured with a motion database, and the motion database stores a plurality of motion training samples; The motion tracking strategy includes a motion tracking method, which includes acquiring the fitter's motion in real time through a stereoscopic vision acquisition subsystem and performing correlation calculations on it with multiple motion training samples in the motion database, and selecting the motion training sample with the highest correlation to display it on a three-dimensional image display module.

7. The virtual fitting method according to claim 5, wherein: The human-computer interaction module includes a walking vibration simulation module, a wind field simulation module and a clothing selection module. The clothing selection module includes clothing classifications of different materials and types. The walking vibration simulation module is used to simulate the vibration of the chest when walking when the selected clothing type is women's clothing. The walking vibration simulation module is configured with a vibration simulation strategy, which includes changing the shape of the three-dimensional model according to the weight, chest circumference and stride of the person trying on the clothes. The wind field simulation module is used to generate wind fields of different directions and different wind speeds.

8. The virtual fitting method according to claim 7, wherein: The three-dimensional image display module of the transparent display screen is arranged on one side or both sides of the transparent display screen, and the transparent display screen includes one or more of a transparent LED screen, an LCD screen, and an OLED screen.

9. An intelligent fitting system for executing a virtual fitting method according to any one of claims 1 to 8, characterized in that: include: The body measurement subsystem includes a body measurement sensor for measuring the body information of the person being fitted and generating weight parameters and body parameters, wherein the body parameters include body fat percentage data and muscle mass data; A stereoscopic vision acquisition subsystem, including a binocular depth camera, is used to collect body image information of the person being fitted and generate body parameters and height parameters, wherein the body parameters include chest circumference data, waist circumference data, hip circumference data, and shoulder width data; A transparent display device, comprising a control and computing subsystem and a transparent display screen electrically connected thereto, wherein the control and computing subsystem is used to obtain the body shape parameters and height parameters and calculate and generate a three-dimensional image of a human body model, and the transparent display screen comprises a human-computer interaction module and a three-dimensional image display module; The control and calculation subsystem is configured with a model improvement strategy, which includes using the weight parameters and physique parameters to calculate and correct the three-dimensional image of the human body model to generate a standard three-dimensional image of the human body model; the control and calculation subsystem is configured with a fitting database, which stores a number of three-dimensional images of clothing of different styles and sizes.

Images