A method and system for creating virtual three-dimensional images

Through virtual 3D image creation methods and systems, the original data is captured in real time, preset models are retrieved and fitted optimization is solved, and the problems of high hardware requirements and long production cycles are achieved in high-precision 3D image creation are solved, and high-quality images are achieved. Fast generation and low-cost production, suitable for users without technical foundations.

CN112381949BActive Publication Date: 2024-11-01HENGXIN SHAMBALA CULTURE CO LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202011211744.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-11-03
Publication Date
2024-11-01
Estimated Expiration
2040-11-03

AI Technical Summary

Technical Problem

In the creation of high-precision three-dimensional images, the hardware requirements are high, the production cycle is long, and users without technical foundations find it difficult to independently complete model construction and image production.

Method used

Provide a virtual three-dimensional image creation method and system, which can obtain the original shooting data in real time, retrieve preset models, perform bonding and optimization processing, and realize the rapid generation and output of three-dimensional images. It uses the acquisition system and the processing system to run in parallel, simplifying User operation.

Benefits of technology

Generating high-quality three-dimensional images under low hardware requirements shortens production cycles and reduces costs, and does not require users to have technical knowledge of three-dimensional image production.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112381949B_ABST
    Figure CN112381949B_ABST
Patent Text Reader

Abstract

The present application discloses a method and system for creating a virtual three-dimensional image, wherein the method comprises the following steps: obtaining at least one original shooting data in real time; calling a corresponding preset model according to the original shooting data; processing the original shooting data and the preset model to obtain a sub-model; synthesizing all the sub-models to obtain a three-dimensional image and output it. The present application can quickly collect the original shooting data of the user and process the original shooting data in real time on the basis of limited hardware equipment, so as to complete and output a short-duration high-quality three-dimensional image in a short time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method and system for creating a virtual three-dimensional image. Background Art

[0002] Reconstructing models with images is one of the relatively new technologies and can be used for some static, low-precision model reconstruction. However, if higher precision is required, a camera matrix is ​​required. If a large scene is to be reconstructed, traditional laser or structured light scanning is required. When processing image data, the carrier hardware of the processing program also has a certain threshold. The overall requirements for device hardware are relatively high.

[0003] The process from collecting user image data to completing the image is a linear one. If the previous task is not completed, the subsequent task cannot be carried out, which leads to an increase in the overall production cycle and time cost.

[0004] The model building process and the production of 3D videos require some related production software, which is difficult for people without technical foundation to complete independently. Summary of the invention

[0005] The purpose of this application is to provide a virtual three-dimensional image creation method and system, which can quickly collect the user's original shooting data and process the original shooting data in real time based on limited hardware equipment, so as to complete and output a short-duration high-quality three-dimensional image in a short time.

[0006] To achieve the above-mentioned purpose, the present application provides a method for creating a virtual three-dimensional image, comprising the following steps: acquiring at least one original shooting data in real time; retrieving a corresponding preset model based on the original shooting data; processing the original shooting data and the preset model to obtain a sub-model; synthesizing all sub-models to obtain a three-dimensional image and output it.

[0007] As above, the sub-steps of calling the corresponding preset model according to the original shooting data are as follows: preprocessing the original shooting data to obtain the shooting angle; generating a model calling instruction according to the shooting angle, and issuing the model calling instruction; receiving the preset model fed back according to the model calling instruction.

[0008] As above, the model calling instruction includes: preset model angle; the expression of preset model angle is as follows: Tjd ys =Tjd ymx ±λ·Tjd zh Among them, Tjd zh =[μ1+μ2·T y (A)]·[μ3+μ4·T x (B)]; where Tjd ysis the preset model angle; Tjd ymx The model angle of the stored model; Tjd zh is the shooting angle; T y (A) is the matrix of deflection angle A; T x (B) is the matrix of the pitch angle B; λ is the adjustment parameter; A is the deflection angle; B is the pitch angle; μ1, μ2, μ3 and μ4 are multiplication coefficients. When there is no deflection angle but a pitch angle in the shooting angle, μ1=1, μ2=0, μ3=0, μ4=1; when there is a deflection angle but no pitch angle in the conversion angle, μ1=0, μ2=1, μ3=1, μ4=0; when there is a deflection angle and a pitch angle in the conversion angle, μ1=0, μ2=1, μ3=0, μ4=1.

[0009] As above, the sub-steps of processing the original shooting data and the preset model to obtain the sub-model are as follows: fitting the original shooting data and the preset model to obtain the fitted model; optimizing the fitted model to obtain the sub-model.

[0010] As above, the sub-steps of fitting the original shooting data and the preset model to obtain the fitted model are as follows: pre-analyze the preset model to obtain the preset head size of the preset model; pre-analyze the original shooting data to obtain multiple original feature points of the original shooting data; pre-adjust the original head size of the head in the original shooting data according to the preset head size and the multiple original feature points to obtain an equal length ratio head; re-process the equal length ratio head and the head in the preset model to obtain a fitted model.

[0011] The present application also provides a virtual three-dimensional image creation system, including: an acquisition system, a control system, a processing system and a storage system; wherein the acquisition system is used to execute data acquisition instructions, obtain original shooting data, and send the original shooting data to the processing system; the control system is used to send data acquisition instructions to the acquisition system; the processing system is used to execute the above-mentioned virtual three-dimensional image creation method; the storage system is used to store storage models; and is used to receive model retrieval instructions, and adjust the storage model according to the model retrieval instructions to obtain a preset model, and feed the preset model back to the processing system.

[0012] As mentioned above, the acquisition system includes: at least three cameras, and the setting position of each camera is different.

[0013] As above, the processing system includes: an analysis device, a model creation device and a data synthesis device; wherein the analysis device is used to process the original shooting data to obtain a preset model, and send the preset model and the original shooting data to the model creation device; the model creation device is used to receive the preset model and the original shooting data; process the preset model and the original shooting data to obtain at least one sub-model, and send the sub-model to the data synthesis device; the data synthesis device is used to synthesize all sub-models to obtain a three-dimensional image and output it.

[0014] As above, the analysis device includes: an identification unit and an analysis unit; wherein the identification unit is used to identify the original shooting data, obtain the avatar area, and send the avatar area to the analysis unit; the analysis unit is used to analyze the avatar area, generate a model call instruction, send the model call instruction to the storage system, and receive a preset model fed back according to the model call instruction; send the preset model and the original shooting data to the model creation device.

[0015] As above, the model creation device includes: a preset recognition model, a preset adjustment model and a preset optimization model; wherein, the preset recognition model is used to process the original shooting data and the preset model to obtain an equal-length ratio head, and input the equal-length ratio head into the preset adjustment model; the preset adjustment model is used to reprocess the equal-length ratio head and the head in the preset model to obtain a fitting model; the preset optimization model is used to optimize the fitting model to obtain a sub-model, and send the sub-model to the data synthesis device.

[0016] The beneficial effects achieved by this application are as follows:

[0017] (1) The virtual three-dimensional image creation method and system of the present application can achieve high-quality three-dimensional images that can meet general needs and have high precision through lower hardware requirements and software processing.

[0018] (2) The virtual three-dimensional image creation method and system of the present application adopts a parallel operation mode of the acquisition system and the processing system. By processing the acquired original shooting data in real time, it replaces the production mode of calculating only after all the original shooting data are collected. This greatly shortens the long production cycle from data acquisition to the completion of the three-dimensional image, and has the advantage of cost saving.

[0019] (3) The virtual three-dimensional image creation method and system of the present application have a user-friendly control system, through which the user can simply and efficiently collect the original shooting data and obtain feedback in real time. The interaction between various systems is carried out through the control system. After the original shooting data is collected, the original shooting data can be automatically processed without manual operation by the user, and the user does not need to have any technical knowledge of building three-dimensional images. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0021] Figure 1 A schematic diagram of the structure of an embodiment of a virtual three-dimensional image creation system;

[0022] Figure 2 A flowchart of an embodiment of a method for creating a virtual three-dimensional image. DETAILED DESCRIPTION

[0023] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] The present application provides a virtual three-dimensional image creation method and system, which can quickly collect the user's original shooting data and process the original shooting data in real time based on limited hardware equipment, so as to complete and output a short-duration high-quality three-dimensional image in a short time.

[0025] like Figure 1 As shown, the present application provides a virtual three-dimensional image creation system, including: an acquisition system 110 , a control system 120 , a processing system 130 and a storage system 140 .

[0026] The acquisition system 110 is used to execute data acquisition instructions, obtain original shooting data, and send the original shooting data to the processing system.

[0027] Control system 120: used to send data collection instructions to the collection system.

[0028] The processing system 130 is used to execute the following virtual three-dimensional image creation method.

[0029] Storage system 140: used to store the storage model; used to receive the model call instruction, and adjust the storage model according to the model call instruction, obtain the preset model, and feed the preset model back to the processing system.

[0030] Furthermore, the acquisition system 110 includes: at least three cameras, and the setting position of each camera is different.

[0031] Specifically, as an embodiment, the camera is a high-definition camera, which is used to collect original data shot by the user.

[0032] Further, the processing system 130 includes: an analysis device, a model creation device, and a data synthesis device;

[0033] Among them, the analysis device is used to process the original shooting data to obtain a preset model, and send the preset model and the original shooting data to the model creation device.

[0034] Model creation device: used for receiving a preset model and original shooting data; processing the preset model and the original shooting data to obtain at least one sub-model, and sending the sub-model to the data synthesis device.

[0035] Data synthesis device: used to synthesize all sub-models to obtain three-dimensional images and output them.

[0036] Furthermore, the analysis device includes: an identification unit and an analysis unit.

[0037] Among them, the recognition unit is used to recognize the original shooting data, obtain the head portrait area, and send the head portrait area to the analysis unit.

[0038] Analysis unit: used to analyze the avatar area, generate a model retrieval instruction, send the model retrieval instruction to the storage system, and receive a preset model fed back according to the model retrieval instruction; send the preset model and the original shooting data to the model creation device.

[0039] Furthermore, the model creation device includes: a preset recognition model, a preset adjustment model and a preset optimization model.

[0040] Among them, the preset recognition model is used to process the original shooting data and the preset model to obtain the equal length ratio head, and input the equal length ratio head into the preset adjustment model.

[0041] Preset adjustment model: used to reprocess the heads with equal length ratio and the heads in the preset model to obtain a fitting model.

[0042] Preset optimization model: used to optimize the fitting model, obtain the sub-model, and send the sub-model to the data synthesis device.

[0043] Furthermore, the control system, the processing system and the storage system are arranged in a computer, and the computer can be connected to two high-definition monitors, which are used to display the creation process of the virtual three-dimensional image creation method and output the three-dimensional image respectively.

[0044] like Figure 2 As shown, the present application provides a method for creating a virtual three-dimensional image, comprising the following sub-steps:

[0045] S210: Acquire at least one original shooting data in real time.

[0046] Furthermore, the sub-steps of obtaining multiple original shooting data in real time are as follows:

[0047] S2101: Receive data collection instructions.

[0048] Specifically, the acquisition system receives the data acquisition instruction sent by the control system and executes S2102.

[0049] S2102: Execute a data acquisition instruction to obtain a plurality of original shooting data.

[0050] Specifically, the acquisition system executes the data acquisition instruction to acquire the original shooting data, and sends the acquired multiple original shooting data to the processing system to execute S220.

[0051] Furthermore, when the acquisition system acquires a piece of original shooting data, the original shooting data can be immediately sent to the processing system for processing, without waiting until all the original shooting data are acquired before sending them to the processing system for processing.

[0052] S220: Retrieve a corresponding preset model according to the original shooting data.

[0053] Furthermore, the sub-steps of retrieving the corresponding preset model according to the original shooting data are as follows:

[0054] S2201: Pre-process the original shooting data to obtain the shooting angle.

[0055] Furthermore, the sub-steps of pre-processing the original shooting data and obtaining the shooting angle are as follows:

[0056] S22011: Identify the shooting data and obtain the portrait area.

[0057] Specifically, after the processing system receives the original shooting data, the recognition unit performs feature recognition on the original shooting data, takes the area with the person's head portrait in the original shooting data as the head portrait area, and sends the head portrait area to the analysis unit.

[0058] S22012: Analyze the head portrait area and obtain the shooting angle.

[0059] Specifically, the analyzing unit analyzes the head portrait area and obtains the angle of the head portrait of the person in the head portrait area as the shooting angle.

[0060] The shooting angles include: deflection angle and pitch angle.

[0061] Specifically, the coordinate axis is set with the original feature point of the nose base of the front face of the person portrait as the origin (0,0). The deflection angle is the angle of rotation of the front face of the person portrait around the Y axis. The pitch angle is the angle of rotation of the front face of the person portrait around the X axis.

[0062] S2202: Generate a model retrieval instruction according to the shooting angle, and issue the model retrieval instruction.

[0063] Specifically, the analysis unit generates a model retrieval instruction according to the shooting angle, and sends the model retrieval instruction to the storage system.

[0064] S2203: Receive the preset model fed back according to the model retrieval instruction.

[0065] Specifically, the model retrieval instruction includes: a preset model angle. After receiving the model retrieval instruction, the storage system adjusts the storage model stored in the storage system according to the preset model angle in the model retrieval instruction, takes the adjusted storage model as the preset model, and feeds the preset model back to the analysis unit. After obtaining the preset model, the analysis unit feeds the preset model and the original shooting data back to the model creation device.

[0066] Furthermore, taking the nose base point (0,0,0) as the coordinate origin and the nose tip point (0,0,h), the expression of the preset model angle is as follows:

[0067] Tj ys =Tjd ymx ±λ·Tjd zh ;

[0068] in,

[0069]

[0070] in,

[0071]

[0072] Among them, Tjd ys is the preset model angle; Tjd ymx The model angle of the stored model; Tjd zh is the shooting angle; T y (A) is the matrix of deflection angle A; T x(B) is the matrix of pitch angle B; λ is the adjustment parameter; A is the deflection angle; B is the pitch angle; h is the distance from the nose tip to the nose base; (a x ,a y ) are the original feature points in the derotated face; μ1, μ2, μ3 and μ4 are multiplication coefficients. When there is no deflection angle but a pitch angle in the shooting angle, μ1=1, μ2=0, μ3=0, μ4=1; when there is a deflection angle but no pitch angle in the conversion angle, μ1=0, μ2=1, μ3=1, μ4=0; when there is a deflection angle and a pitch angle in the conversion angle, μ1=0, μ2=1, μ3=0, μ4=1.

[0073] Specifically, when Tjd ymx >Tjd zh When Tjd ys =Tjd ymx -λ·Tjd zh ; When Tjd ymx <Tjd zh When Tjd ys =Tjd ymx +λ·Tjd zh .

[0074] S230: Processing the original shooting data and the preset model to obtain a sub-model.

[0075] Furthermore, the sub-steps of processing the original shooting data and the preset model to obtain the sub-model are as follows:

[0076] S2301: Perform fitting processing on the original shooting data and the preset model to obtain a fitting model.

[0077] Specifically, the sub-steps of fitting the original shooting data and the preset model to obtain the fitted model are as follows:

[0078] S23011: Pre-analyze the preset model to obtain a preset head size of the preset model.

[0079] Specifically, after the model creation device obtains the original shooting data and the preset model, it identifies the preset model through a preset recognition model, and determines multiple calibration points of the head in the preset model, where the calibration points include: forehead calibration point, jaw calibration point, left face contour calibration point, right face contour calibration point, eye calibration point, nose calibration point, ear calibration point and mouth calibration point, etc.

[0080] The preset recognition model determines the preset head size of the preset model according to the multiple calibration points. The preset head size includes at least: a preset head length and a preset head width. The preset head length is the distance between the highest calibration point in the forehead calibration points and the lowest calibration point in the chin calibration points. The preset head width is the distance between the leftmost calibration point in the left face contour calibration points and the rightmost calibration point in the right face contour calibration points.

[0081] S23012: Pre-analyze the original shooting data to obtain multiple original feature points of the original shooting data.

[0082] Specifically, after obtaining the preset head size, the model creation device identifies the preset model through a preset recognition model, and determines a plurality of original feature points of the head in the original captured data, wherein the number of original feature points is the same as the number of calibration points. Each original feature point has an original feature point name, each calibration point has a calibration point name, and the original feature point name corresponds to the calibration point name one by one. For example, the original feature points include: forehead original feature points, jaw original feature points, left face contour original feature points, right face contour original feature points, eye original feature points, nose original feature points, ear original feature points, and mouth original feature points.

[0083] S23013: Pre-adjusting the original head size of the head in the original captured data according to a preset head size and a plurality of original feature points to obtain a head with an equal length ratio.

[0084] Specifically, the original head size includes at least: the original head length and the original head width. The original head size is the distance between the original feature point at the highest position in the original feature points of the forehead and the original feature point at the lowest position in the original feature points of the jaw. The original head size is the distance between the original feature point at the leftmost position in the original feature points of the left face contour and the original feature point at the rightmost position in the original feature points of the right face contour.

[0085] The preset recognition model uses the preset head length and the original feature points in the preset head size to perform an overall scaling adjustment on the original head size of the head in the original shooting data, so that the original head length is equal to the preset head length, but does not affect the ratio of the original head width to the original head length. The head in the original shooting data after the overall scaling adjustment is taken as an equal length ratio head, and the equal length ratio head is input into the preset adjustment model.

[0086] S23014: Reprocess the head with equal length ratio and the head in the preset model to obtain a fitting model.

[0087] Furthermore, the head with equal length ratio and the head in the preset model are reprocessed to obtain the fitting model in the following sub-steps:

[0088] F1: Reprocess the equal-length-ratio head to obtain an equal-length-ratio mesh head.

[0089] Specifically, the preset adjustment model uses all original feature points in the equal-ratio head to perform grid processing on the equal-ratio head to obtain an equal-ratio grid avatar.

[0090] F2: Reprocess the head in the preset model to obtain the preset model mesh head.

[0091] Specifically, the preset adjustment model uses all the multiple calibration points to perform grid processing on the head in the preset model to obtain the preset model grid head.

[0092] F3: Use the equal-length mesh head to adjust the preset model mesh head to obtain a fitting model.

[0093] Specifically, the preset adjustment model uses an equal-length ratio grid head to adjust the preset model grid head, so that the preset head width of the preset model grid head after adjustment is equal to the head width of the equal-length ratio head, and then adjusts the corresponding calibration points according to the original feature points in the equal-length ratio grid head, and fits the equal-length ratio grid head with the adjusted preset model grid head to obtain a fitted model.

[0094] S2302: Optimize the fitting model to obtain a sub-model.

[0095] Specifically, the preset adjustment model sends the fitting model to the preset optimization model for optimization processing, and uses the optimized fitting model as a sub-model, and sends the sub-model to the data synthesis device. Optimizing the fitting model can effectively reduce image crosstalk.

[0096] Furthermore, the sub-model is input into the optimization model for pixel correction, wherein the expression of the corrected pixel is as follows:

[0097]

[0098] Among them, G xz is the corrected pixel; g r is the preset target light intensity information of the simulated right eye; g l is the preset target light intensity information for simulating the left eye; κ1, κ2, κ3 and κ4 are four sub-information for simulating the pixel light intensity of the right eye; η1, η2, η3 and η4 are four sub-information for simulating the pixel light intensity of the left eye; υ is a parameter, υ<κ1+κ2+κ3+κ4 and υ<η1+η2+η3+η4.

[0099] S240: Perform synthesis processing on all sub-models to obtain and output a three-dimensional image.

[0100] Furthermore, the sub-steps of synthesizing all sub-models to obtain and output a three-dimensional image are as follows:

[0101] S2401: Combine all sub-models into one audio-visual data.

[0102] Specifically, after receiving all the sub-models, the data synthesis device synthesizes all the sub-models into one audio-visual data.

[0103] S2402: Select at least one background data and at least one audio data, and synthesize the audio and video data, the background data and the audio data into a three-dimensional image, and output the three-dimensional image.

[0104] Specifically, the data synthesis device automatically selects or receives at least one background data and at least one audio data selected by the user, and synthesizes the audio and video data, the background data and the audio data into a three-dimensional image, and outputs the three-dimensional image.

[0105] The beneficial effects achieved by this application are as follows:

[0106] (1) The virtual three-dimensional image creation method and system of the present application can achieve high-quality three-dimensional images that can meet general needs and have high precision through lower hardware requirements and software processing.

[0107] (2) The virtual three-dimensional image creation method and system of the present application adopts a parallel operation mode of the acquisition system and the processing system. By processing the acquired original shooting data in real time, it replaces the production mode of calculating only after all the original shooting data are collected. This greatly shortens the long production cycle from data acquisition to the completion of the three-dimensional image, and has the advantage of cost saving.

[0108] (3)) The virtual three-dimensional image creation method and system of the present application have a user-friendly control system, through which the user can simply and efficiently collect the original shooting data and obtain feedback in real time. The interaction between various systems is carried out through the control system. After the original shooting data is collected, the original shooting data can be automatically processed without manual operation by the user, and the user does not need to have any technical knowledge of building three-dimensional images.

[0109] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic creative concepts. Therefore, the scope of protection of the present application is intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application. Obviously, those skilled in the art may make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the protection scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A method for creating a virtual three-dimensional image, characterized in that: The steps include: Acquire at least one piece of original shooting data in real time; Retrieve the corresponding preset model according to the original shooting data; Processing the original shooting data and the preset model to obtain a sub-model; Perform synthesis processing on all sub-models to obtain and output three-dimensional images; The sub-steps of retrieving the corresponding preset model according to the original shooting data are as follows: Pre-process the original shooting data to obtain the shooting angle; Generate model retrieval instructions based on the shooting angle and issue the model retrieval instructions; Receiving a preset model according to a model call instruction feedback; The model calling instruction includes: preset model angle; the expression of preset model angle is as follows: Time ys =Time ymx ±λ·Tjd zh ; in, Tjd zh =[μ1+μ2·T y (A)]·[μ3+μ4·T x (B)]; Among them, Tjd ys is the preset model angle; Tjd ymx The model angle of the stored model; Tjd zh is the shooting angle; T y (A) is the matrix of deflection angle A; T x (B) is the matrix of the pitch angle B; λ is the adjustment parameter; A is the deflection angle; B is the pitch angle; μ1, μ2, μ3 and μ4 are multiplication coefficients. When there is no deflection angle but a pitch angle in the shooting angle, μ1=1, μ2=0, μ3=0, μ4=1; when there is a deflection angle but no pitch angle in the conversion angle, μ1=0, μ2=1, μ3=1, μ4=0; when there is a deflection angle and a pitch angle in the conversion angle, μ1=0, μ2=1, μ3=0, μ4=1.

2. The method for creating a virtual three-dimensional image according to claim 1, characterized in that: The sub-steps of processing the original shooting data and the preset model to obtain the sub-model are as follows: The original shooting data and the preset model are fitted to obtain a fitted model; The fitted model is optimized to obtain a sub-model.

3. The method for creating a virtual three-dimensional image according to claim 2, characterized in that: The sub-steps of shooting the original data and the preset model for fitting are as follows: Pre-analyzing the preset model to obtain a preset head size of the preset model; Pre-analyze the original shooting data to obtain multiple original feature points of the original shooting data; Pre-adjusting the original head size of the head in the original captured data according to the preset head size and multiple original feature points to obtain a head with equal length ratio; The head with equal length ratio and the head in the preset model are reprocessed to obtain a fitting model.

4. A virtual three-dimensional image creation system, characterized in that: include: Acquisition system, control system, processing system and storage system; The acquisition system is used to execute data acquisition instructions, obtain the original shooting data, and send the original shooting data to the processing system; Control system: used to send data collection instructions to the collection system; Processing system: used to execute the virtual three-dimensional image creation method as described in any one of claims 1 to 3; Storage system: used to store the storage model; used to receive the model call instruction, and adjust the storage model according to the model call instruction, obtain the preset model, and feed the preset model back to the processing system.

5. The virtual three-dimensional image creation system according to claim 4, characterized in that: The acquisition system includes: at least three cameras, and the setting position of each camera is different.

6. The virtual three-dimensional image creation system according to claim 4, characterized in that: The processing system includes: an analysis device, a model creation device and a data synthesis device; The analyzing device is used to process the original shooting data to obtain a preset model, and send the preset model and the original shooting data to the model creating device; Model creation device: used for receiving a preset model and original shooting data; processing the preset model and original shooting data to obtain at least one sub-model, and sending the sub-model to the data synthesis device; Data synthesis device: used to synthesize all sub-models to obtain three-dimensional images and output them.

7. The virtual three-dimensional image creation system according to claim 6, characterized in that: The analysis device comprises: an identification unit and an analysis unit; The recognition unit is used to recognize the original photographed data, obtain the head portrait area, and send the head portrait area to the analysis unit; Analysis unit: used to analyze the avatar area, generate a model retrieval instruction, send the model retrieval instruction to the storage system, and receive a preset model fed back according to the model retrieval instruction; send the preset model and the original shooting data to the model creation device.

8. The virtual three-dimensional image creation system according to claim 6, characterized in that: The model creation device includes: a preset recognition model, a preset adjustment model and a preset optimization model; The preset recognition model is used to process the original shooting data and the preset model to obtain the equal-length ratio head, and input the equal-length ratio head into the preset adjustment model; Preset adjustment model: used to reprocess the head with equal length ratio and the head in the preset model to obtain a fitting model; Preset optimization model: used to optimize the fitting model, obtain the sub-model, and send the sub-model to the data synthesis device.

Citation Information

Patent Citations

  • Virtual three-dimensional human body modeling method and modeling system

    CN106250573A

  • Image processing method and device, storage medium and electronic equipment

    CN108495032A

  • Clothing try-on method and device, storage medium and mobile terminal

    CN109978640A