Cloud design cooperation platform system based on customized model training

Abstract

The invention discloses a cloud design collaboration platform system based on customized model training, relates to a cloud collaboration domain, solves the problem that the accuracy of costume design cannot be guaranteed by the current costume cloud collaboration design, and comprises a costume segmentation module, a color analysis module, a texture analysis module and a comprehensive analysis module, the costume segmentation module is used for segmenting the costume design drawing to obtain a plurality of costume design areas corresponding to the costume design drawing; the color analysis module is used for analyzing the color of the costume design area to obtain an abnormal signal or a color excellent value of the costume design area; the texture analysis module is used for analyzing the texture of the costume design area to obtain an abnormal signal or a texture similarity level of the costume design area; the comprehensive analysis module is used for analyzing the comprehensive condition of the costume design drawing to obtain an entity costume drawing corresponding to the costume design drawing or generate a modification signal, and accurate design of the costume is achieved based on the costume design drawing.

Description

Technical Field

[0001] The present invention relates to cloud collaboration technology, and in particular to a cloud design collaboration platform system based on customized model training. Background Art

[0002] With the development of society, people's requirements for clothing are getting higher and higher. However, when designing clothing, it is necessary to combine market demand, fabric selection, color matching and visual effects and other factors. Therefore, designing a new piece of clothing is very time-consuming. Therefore, cloud-based design collaboration platforms have emerged. Cloud-based design platforms on the market allow multiple users to design the same clothing, and each user is only responsible for the part they are involved in. With the development of technology, cloud-based clothing design may continue to develop intelligence, personalization and integration with the real world, providing a more intelligent, interactive and personalized design experience.

[0003] However, at present, when designing clothing in the cloud, most processes require real-time supervision by users. If users cannot supervise in real time, they cannot judge whether the clothing design meets the user's requirements based on the differences in color and texture in the clothing design, and thus cannot achieve accurate clothing design.

[0004] To this end, we propose a cloud-based design collaboration platform system based on customized model training. Summary of the invention

[0005] The purpose of the present invention is to propose a cloud-based design collaboration platform system based on customized model training to solve the problem in the above-mentioned background technology that it is impossible to accurately design clothing through color differences and texture differences.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A cloud-based design collaboration platform system based on customized model training, including a data acquisition module, a clothing generation module, a clothing segmentation module, a color analysis module, a texture analysis module, a comprehensive analysis module, a database, and a user terminal;

[0008] The data collection module is used to collect the past clothing images and design requirements corresponding to the prepared clothing designs input by the user, and send the past clothing images and design requirements to the clothing generation module;

[0009] The clothing generation module is used to draw a clothing design drawing corresponding to the prepared clothing design according to the past clothing images and the design requirements of the prepared clothing design, and send the clothing design drawing to the clothing segmentation module;

[0010] The database is used to store a plurality of texture images used for generating clothing design drawings, and send the plurality of texture images used for the clothing design drawings to the comprehensive analysis module;

[0011] The clothing segmentation module is used to segment the clothing design drawing, obtain multiple clothing design areas corresponding to the clothing design drawing and send them to the color analysis module and the texture analysis module;

[0012] The color analysis module is used to analyze the color of the clothing design area of ​​the clothing design drawing. If the analysis obtains a color excellence value corresponding to the clothing design area, the color analysis module sends the color excellence value corresponding to the clothing design drawing to the comprehensive analysis module. If the analysis generates an abnormal signal, the color analysis module sends the abnormal signal to the user terminal.

[0013] The texture analysis module is used to analyze the texture of the clothing design drawing, and the abnormal signal obtained by the analysis is sent to the user terminal, or the texture similarity level of multiple clothing design areas is analyzed and sent to the comprehensive analysis module;

[0014] The comprehensive analysis module is used to analyze the comprehensive situation of the clothing design drawing, analyze the physical clothing picture corresponding to the clothing design drawing, or analyze and obtain a modification signal to send to the user terminal;

[0015] The user terminal is used to receive an abnormal signal or a modification signal. If it is an abnormal signal, the clothing generation module is re-executed to generate a new preliminary design clothing picture; if it is a modification signal, the corresponding clothing design area is modified.

[0016] Furthermore, the design requirements are clothing tendencies and color tendencies of the clothing to be designed, and the clothing tendencies include tops, bottoms, and jumpsuits.

[0017] Furthermore, the segmentation process of the clothing segmentation module is as follows:

[0018] Obtain a clothing design drawing, and obtain clothing trends corresponding to the clothing design drawing;

[0019] If the clothing tendency includes a top, the clothing design drawing is divided according to the sleeve corresponding area, the neckline corresponding area, the front piece corresponding area, and the back piece corresponding area, and recorded as the first clothing design area, the second clothing design area, the third clothing design area, and the fourth clothing design area in sequence;

[0020] If the clothing tendency includes bottoms, the clothing design drawing is divided into the corresponding areas of the trouser legs, the corresponding areas of the trouser legs and the corresponding areas of the crotch, and are respectively recorded as the fifth clothing design area, the sixth clothing design area and the seventh clothing design area.

[0021] Furthermore, the analysis process of the color analysis module is as follows:

[0022] Acquire the color tendency corresponding to the prepared design clothing in the first clothing design area, and obtain the preset color RGB values ​​of different colors in the first clothing design area;

[0023] Obtaining the RGB value of each pixel corresponding to the first clothing design area, counting the number of pixels corresponding to the preset color RGB values ​​of different colors in the first clothing area and dividing the number by the total number of pixels in the first clothing design area, to obtain the first color proportion of different colors in the first clothing area;

[0024] If the first color proportion of any color is not equal to the corresponding standard color proportion, an abnormal signal is generated and no subsequent process is performed;

[0025] If the first color proportions of all colors are equal to the corresponding standard color proportions, proceed to the subsequent steps.

[0026] Furthermore, the analysis process of the color analysis module also includes:

[0027] Calculate the red first-order color moment and the red second-order color moment of the corresponding color in the first clothing design area;

[0028] The first-order color moment of red is: Among them, o is the color number, i represents the number of the pixel in the first clothing design area, Rio represents the R value component of the RGB value corresponding to the pixel, and Uro represents the color deviation of the corresponding clothing design area. N is the upper limit value of i, and N is an Arabic numeral;

[0029] The second-order color moment is: The red first-order color moment is used to characterize the average response intensity of the red component in different colors in the first clothing design area, and the red second-order color moment is used to represent the response variance of the red component in different colors in the first clothing design area;

[0030] The red excellence values ​​RYXo of different colors in the first clothing design area are calculated by the formula, and the specific formula is as follows:

[0031] RYXo=1 / (A1×YJro+A2×EJro); where A1 and A2 are weight coefficients with fixed values, and A2>A1;

[0032] Similarly, calculate the green excellence value and blue excellence value of different colors in the first clothing design area;

[0033] The red excellent value, green excellent value and blue excellent value of different colors in the first clothing design area are calculated by the formula to obtain the color excellent value of all colors in the first clothing design area. The formula is:

[0034] Color excellent value = red excellent value * weight coefficient 1 + green excellent value * weight coefficient 2 + blue excellent value * weight coefficient 3;

[0035] Traverse and compare the color excellence values ​​of all colors in the first clothing design area, obtain the minimum value among the color excellence values, and use the minimum value as the first color excellence value of the first clothing design area;

[0036] Similarly, the second to seventh color excellence values ​​corresponding to the second to seventh clothing design areas are obtained.

[0037] Furthermore, the analysis process of the texture analysis module is specifically as follows:

[0038] Obtain the RGB value of the pixel point in the first clothing design area, and add the R value component, the G value component, and the B value component in the RGB value of the same pixel point to obtain the sum of the R value component, the G value component, and the B value component as the pixel value XSZi of the corresponding pixel point;

[0039] Take every nine pixels as an independent window, record the surrounding eight pixels as edge pixels, and record the middle pixel as the center pixel; start with the edge pixel in the upper left corner, rotate clockwise, and compare the pixel values ​​of the edge pixels with the pixel values ​​of the center pixel;

[0040] If the pixel value of the edge pixel is greater than or equal to the pixel value of the center pixel, the gray value of the edge pixel is set to 1; if the pixel value of the edge pixel is less than the pixel value of the center pixel, the gray value of the edge pixel is set to 0;

[0041] The pixel points with a gray value of 1 are connected to obtain a preliminary texture image corresponding to the first clothing design area.

[0042] Furthermore, the analysis process of the texture analysis module also includes:

[0043] Calculate the Hamming distance between the prepared texture image and multiple texture images preset in the database by using an image hash algorithm;

[0044] If the Hamming distance of the first clothing design area is greater than or equal to the first Hamming threshold, an abnormal signal is generated;

[0045] If the Hamming distance of the first clothing design region is less than the first Hamming threshold and greater than or equal to the second Hamming threshold, the texture similarity level of the first clothing design region is recorded as first-level similarity;

[0046] If the Hamming distance of the first clothing design area is less than the second Hamming threshold, the texture similarity level of the first clothing design area is recorded as secondary similarity; wherein the first Hamming threshold is greater than the second Hamming threshold, and the texture similarity corresponding to the primary similarity is less than the texture similarity corresponding to the secondary similarity;

[0047] Similarly, the texture similarity levels of the second clothing design area to the seventh clothing design area are obtained.

[0048] Furthermore, the analysis process of the comprehensive analysis module is as follows:

[0049] Obtain the first color excellent value SC1 to the seventh color excellent value SC7 corresponding to the clothing design drawing;

[0050] Obtaining texture similarity levels of the first clothing design region to the seventh clothing design region;

[0051] Calculate the clothing excellence value of the first clothing design area, FY1 = SC1 × λ, where λ is the texture similarity weight value, and the texture similarity weight value corresponding to the first-level similarity is less than the texture similarity weight value corresponding to the second-level similarity;

[0052] Similarly, the clothing excellence values ​​of the second clothing design area to the seventh clothing design area are calculated;

[0053] Compare the clothing excellence values ​​of the first to seventh clothing design areas with the clothing excellence threshold; if the clothing excellence values ​​of all clothing design areas are greater than the clothing excellence threshold, automatically generate a physical clothing image corresponding to the clothing design image through the artificial intelligence model;

[0054] If the clothing excellence value of any of the first to seventh clothing design areas is less than the clothing excellence threshold, a modification signal for the corresponding clothing design area is generated.

[0055] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0056] The present invention first collects the past clothing images input by the user and the design requirements corresponding to the prepared clothing designs. The clothing generation module draws the clothing design drawings corresponding to the prepared clothing designs according to the past clothing images and the design requirements of the prepared clothing designs. The clothing segmentation module segments the clothing design drawings to obtain multiple clothing design areas corresponding to the clothing design drawings. The color analysis module is then used to analyze the colors of the clothing design areas of the clothing design drawings. If the analysis obtains the color excellence value corresponding to the clothing design area, it is sent to the comprehensive analysis module. If the analysis generates an abnormal signal, it is sent to the user terminal. On the other hand, the texture of the clothing design drawings is analyzed by the texture analysis module. If the analysis obtains an abnormal signal, it is sent to the user terminal. If the analysis obtains the texture similarity level of the clothing design area, it is sent to the comprehensive analysis module. Finally, the comprehensive analysis module is used to analyze the comprehensive situation of the clothing design drawings, and the entity clothing drawings corresponding to the clothing design drawings are analyzed or the modification signals are obtained. The present invention realizes accurate design of clothing based on the clothing design drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0058] Figure 1 It is the overall system block diagram of the present invention.

[0059] Figure 2 It is a schematic diagram of the conversion between a pixel image and a grayscale image in the present invention.

[0060] Figure 3 This is a schematic diagram of preparing designed clothing in the present invention.

[0061] Figure 4 The physical clothing image corresponding to the clothing designed in the present invention is prepared. DETAILED DESCRIPTION

[0062] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] Embodiment 1:

[0064] See also Figure 1-Figure 4 As shown, the technical solution provided by the present invention is: a cloud-based design collaboration platform system based on customized model training, including a data acquisition module, a clothing generation module, a clothing segmentation module, a color analysis module, a texture analysis module, a comprehensive analysis module, a database and a user terminal;

[0065] The data collection module is used to collect the past clothing images and design requirements corresponding to the prepared clothing designs input by the user, and send the past clothing images and design requirements to the clothing generation module, wherein the design requirements specifically include the color tendency and clothing tendency (tops, bottoms or jumpsuits) of the prepared clothing designs;

[0066] The clothing generation module is used to draw a clothing design drawing corresponding to the prepared clothing design according to the past clothing images and the design requirements of the prepared clothing design, and send the clothing design drawing to the clothing segmentation module; wherein the clothing generation module specifically obtains a corresponding clothing style training model by training the past clothing images, inputs the design requirements of the prepared clothing design into the clothing style training model, and generates a clothing design drawing corresponding to the prepared clothing design;

[0067] The database is connected to the comprehensive analysis module for data storage, and the database is used to store a plurality of texture images used for generating clothing design drawings, and send the plurality of texture images used for the clothing design drawings to the comprehensive analysis module;

[0068] The clothing segmentation module is used to segment the clothing design drawings generated by the style training model. The segmentation process is as follows:

[0069] Get the clothing design image generated by the style training model, and segment the clothing design image to obtain the clothing design area. The segmentation process is as follows:

[0070] Get the clothing tendency corresponding to the clothing design drawing;

[0071] If the clothing tendency includes a top, the clothing design drawing is divided according to the sleeve corresponding area, the neckline corresponding area, the front piece corresponding area, and the back piece corresponding area, and recorded as the first clothing design area, the second clothing design area, the third clothing design area, and the fourth clothing design area in sequence;

[0072] If the clothing tendency includes bottoms, the clothing design drawing is divided into the trouser leg corresponding area, the trouser leg corresponding area and the crotch corresponding area, and they are recorded as the fifth clothing design area, the sixth clothing design area and the seventh clothing design area respectively;

[0073] The clothing segmentation module sends the clothing design area corresponding to the clothing design drawing to the color analysis module and the texture analysis module;

[0074] It should be specifically noted that if it is only a top, the fifth to seventh clothing design areas are blank, and if it is only a bottom, the first to fourth clothing design areas are blank;

[0075] The color analysis module is used to analyze the color of the clothing design area of ​​the clothing design drawing. The analysis process is as follows:

[0076] Acquire the color tendency corresponding to the prepared design clothing in the first clothing design area, and obtain the preset color RGB values ​​of different colors in the first clothing design area;

[0077] Obtaining the RGB value of each pixel corresponding to the first clothing design area, counting the number of pixels corresponding to the preset color RGB values ​​of different colors in the first clothing area and dividing the number by the total number of pixels in the first clothing design area, to obtain the first color proportion of different colors in the first clothing area;

[0078] If the first color proportion of any color is not equal to the corresponding standard color proportion, an abnormal signal is generated and no subsequent process is performed;

[0079] If the first color proportions of all colors are equal to the corresponding standard color proportions, the red first-order color moment and the red second-order color moment of the corresponding colors in the first clothing design area are calculated;

[0080] Among them, the first-order color moment of red is: Among them, o is the color number, i represents the number of the pixel in the first clothing design area, Rio represents the R value component of the RGB value corresponding to the pixel, and Uro represents the color deviation of the corresponding clothing design area. N is the upper limit value of i, and N is an Arabic numeral;

[0081] The second-order color moment is: The red first-order color moment is used to characterize the average response intensity of the red component in different colors in the first clothing design area, and the red second-order color moment is used to represent the response variance of the red component in different colors in the first clothing design area;

[0082] The red excellence values ​​RYXo of different colors in the first clothing design area are calculated by the formula, and the specific formula is as follows:

[0083] RYXo=1 / (A1×YJro+A2×EJro); where A1 and A2 are weight coefficients with fixed values, and A2>A1;

[0084] Similarly, calculate the green excellence value and blue excellence value of different colors in the first clothing design area;

[0085] The red excellent value, green excellent value and blue excellent value of different colors in the first clothing design area are calculated by the formula to obtain the color excellent value of all colors in the first clothing design area. The formula is:

[0086] Color excellent value = red excellent value * weight coefficient 1 + green excellent value * weight coefficient 2 + blue excellent value * weight coefficient 3;

[0087] Traverse and compare the color excellence values ​​of all colors in the first clothing design area, obtain the minimum value among the color excellence values, and use the minimum value as the first color excellence value of the first clothing design area;

[0088] Similarly, the second color excellence value to the seventh color excellence value corresponding to the second clothing design area to the seventh clothing design area are obtained;

[0089] If the analysis obtains the color excellence value corresponding to the clothing design area, the color analysis module sends the color excellence value corresponding to the clothing design drawing to the comprehensive analysis module; if the analysis generates an abnormal signal, the color analysis module sends the abnormal signal to the user terminal;

[0090] The texture analysis module is used to analyze the texture of the clothing design image generated by the style training model. The analysis process is as follows:

[0091] Obtain the RGB value of the pixel point in the first clothing design area, and add the R value component, the G value component, and the B value component in the RGB value of the same pixel point to obtain the sum of the R value component, the G value component, and the B value component as the pixel value XSZi of the corresponding pixel point;

[0092] like Figure 2 As shown, every nine pixels are taken as an independent window, the surrounding eight pixels are recorded as edge pixels, and the middle pixel is recorded as the center pixel; starting from the edge pixel in the upper left corner, rotate clockwise, and compare the pixel values ​​of the edge pixels with the pixel values ​​of the center pixel respectively;

[0093] If the pixel value of the edge pixel is greater than or equal to the pixel value of the center pixel, the gray value of the edge pixel is set to 1; if the pixel value of the edge pixel is less than the pixel value of the center pixel, the gray value of the edge pixel is set to 0;

[0094] Connect the pixel points with gray value of 1 to obtain the preliminary texture image corresponding to the first clothing design area;

[0095] Calculate the similarity between the prepared texture image and a plurality of texture images preset in the database by using an image hash algorithm;

[0096] Among them, the hash algorithm is an existing technology. The image hash algorithm evaluates the image similarity by converting the image into a hash value of fixed length and then comparing the Hamming distance. Common hash algorithms include average hash algorithm (aHash), perceptual hash algorithm (pHash) and difference hash algorithm (dHash);

[0097] Hash value: Maps a binary value string of arbitrary length to a binary value string of fixed length. The mapping rule is the hash algorithm, and the binary value string obtained after mapping the original data is the hash value.

[0098] Hamming distance: the number of steps required to transform one set of binary data into another set of binary data. This value can measure the difference between two images. The smaller the Hamming distance, the higher the similarity. A Hamming distance of 0 means that the two images are exactly the same.

[0099] Select the minimum value of the Hamming distance as the Hamming distance of the first clothing design area;

[0100] If the Hamming distance of the first clothing design area is greater than or equal to the first Hamming threshold, an abnormal signal is generated;

[0101] If the Hamming distance of the first clothing design region is less than the first Hamming threshold and greater than or equal to the second Hamming threshold, the texture similarity level of the first clothing design region is recorded as first-level similarity;

[0102] If the Hamming distance of the first clothing design area is less than the second Hamming threshold, the texture similarity level of the first clothing design area is recorded as secondary similarity; wherein the first Hamming threshold is greater than the second Hamming threshold, and the texture similarity corresponding to the primary similarity is less than the texture similarity corresponding to the secondary similarity;

[0103] Similarly, the texture similarity levels of the second clothing design area to the seventh clothing design area are obtained;

[0104] The texture analysis module sends the abnormal signal to the user terminal, and sends the texture similarity levels of the first clothing design area to the seventh clothing design area to the comprehensive analysis module;

[0105] The comprehensive analysis module is used to analyze the comprehensive situation of the clothing design drawing. The specific analysis process is as follows:

[0106] Obtain the first color excellent value SC1 to the seventh color excellent value SC7 corresponding to the clothing design drawing;

[0107] Obtaining texture similarity levels of the first clothing design region to the seventh clothing design region;

[0108] Calculate the clothing excellence value of the first clothing design area, FY1 = SC1 × λ, where λ is the texture similarity weight value, and the texture similarity weight value corresponding to the first-level similarity is less than the texture similarity weight value corresponding to the second-level similarity;

[0109] Similarly, the clothing excellence values ​​of the second clothing design area to the seventh clothing design area are calculated;

[0110] Compare the clothing excellence values ​​of the first to seventh clothing design areas with the clothing excellence threshold; if the clothing excellence values ​​of all clothing design areas are greater than the clothing excellence threshold, then Figure 3-Figure 4 As shown, the physical clothing image corresponding to the clothing design image is automatically generated through the artificial intelligence model;

[0111] If the clothing excellence value of any clothing design area from the first to the seventh clothing design area is less than the clothing excellence threshold, a modification signal for the corresponding clothing design area is generated;

[0112] The comprehensive analysis module sends the modification signal to the user terminal;

[0113] The user terminal is used to receive an abnormal signal or a modification signal. If it is an abnormal signal, the clothing generation module is re-executed to generate a new preliminary design clothing image; if it is a modification signal, the corresponding clothing design area is modified;

[0114] In this application, if corresponding calculation formulas appear, the above calculation formulas are all dimensionless and take their numerical calculations. The weight coefficients, proportional coefficients and other coefficients in the formulas are set to a result value obtained by quantifying each parameter. The size of the weight coefficient and the proportional coefficient can be determined as long as it does not affect the proportional relationship between the parameter and the result value.

[0115] Embodiment 2:

[0116] Based on another concept of the same invention, a cloud-based design collaboration method based on customized model training is proposed, comprising the following steps:

[0117] Step S101, drawing a clothing design drawing corresponding to the prepared clothing design according to the past clothing images and the design requirements of the prepared clothing design;

[0118] Step S102, segmenting the clothing design drawing to obtain a plurality of clothing design areas corresponding to the clothing design drawing;

[0119] Step S103, analyzing the color conditions of the clothing design areas to obtain color excellence values ​​of multiple clothing design areas or obtain abnormal signals;

[0120] Step S104, analyzing the texture conditions of the multiple clothing design areas to obtain texture similarity levels of the multiple clothing design areas or obtain abnormal signals;

[0121] Step S105, combining the color excellence value and texture similarity level of the clothing design area, obtaining the physical clothing image corresponding to the clothing design image or obtaining a modification signal and sending it to the user terminal, the user terminal receives the abnormal signal or modification signal and makes a corresponding response.

[0122] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A cloud-based design collaboration platform system based on customized model training, characterized by: It includes data acquisition module, clothing generation module, clothing segmentation module, color analysis module, texture analysis module, comprehensive analysis module, database and user terminal; The data collection module is used to collect the past clothing images and design requirements corresponding to the prepared clothing designs input by the user, and send the past clothing images and design requirements to the clothing generation module; The clothing generation module is used to draw a clothing design drawing corresponding to the prepared clothing design according to the past clothing images and the design requirements of the prepared clothing design, and send the clothing design drawing to the clothing segmentation module; The database is used to store a plurality of texture images used for generating clothing design drawings, and send the plurality of texture images used for the clothing design drawings to the comprehensive analysis module; The clothing segmentation module is used to segment the clothing design drawing, obtain multiple clothing design areas corresponding to the clothing design drawing and send them to the color analysis module and the texture analysis module; The color analysis module is used to analyze the color of the clothing design area of ​​the clothing design drawing. If the analysis obtains a color excellence value corresponding to the clothing design area, the color analysis module sends the color excellence value corresponding to the clothing design drawing to the comprehensive analysis module. If the analysis generates an abnormal signal, the color analysis module sends the abnormal signal to the user terminal. The texture analysis module is used to analyze the texture of the clothing design drawing, and the abnormal signal obtained by the analysis is sent to the user terminal, or the texture similarity level of multiple clothing design areas is analyzed and sent to the comprehensive analysis module; The comprehensive analysis module is used to analyze the comprehensive situation of the clothing design drawing, obtain the physical clothing picture corresponding to the clothing design drawing, or analyze and obtain a modification signal to send to the user terminal; The user terminal is used to receive an abnormal signal or a modification signal. If it is an abnormal signal, the clothing generation module is re-executed to generate a new preliminary design clothing picture; if it is a modification signal, the corresponding clothing design area is modified.

2. The cloud-based design collaboration platform system based on customized model training according to claim 1 is characterized in that: The design requirements are clothing tendencies and color tendencies of the clothing to be designed, and the clothing tendencies include tops, bottoms, and jumpsuits.

3. The cloud-based design collaboration platform system based on customized model training according to claim 1 is characterized in that: The segmentation process of the clothing segmentation module is as follows: Obtain a clothing design drawing, and obtain clothing trends corresponding to the clothing design drawing; If the clothing tendency includes a top, the clothing design drawing is divided according to the sleeve corresponding area, the neckline corresponding area, the front piece corresponding area, and the back piece corresponding area, and recorded as the first clothing design area, the second clothing design area, the third clothing design area, and the fourth clothing design area in sequence; If the clothing tendency includes bottoms, the clothing design drawing is divided into the corresponding areas of the trouser legs, the corresponding areas of the trouser legs and the corresponding areas of the crotch, and are respectively recorded as the fifth clothing design area, the sixth clothing design area and the seventh clothing design area.

4. The cloud-based design collaboration platform system based on customized model training according to claim 1, characterized in that: The analysis process of the color analysis module is as follows: Acquire the color tendency corresponding to the prepared design clothing in the first clothing design area, and obtain the preset color RGB values ​​of different colors in the first clothing design area; Obtaining the RGB value of each pixel corresponding to the first clothing design area, counting the number of pixels corresponding to the preset color RGB values ​​of different colors in the first clothing area and dividing the number by the total number of pixels in the first clothing design area, to obtain the first color proportion of different colors in the first clothing area; If the first color proportion of any color is not equal to the corresponding standard color proportion, an abnormal signal is generated and no subsequent process is performed; If the first color proportions of all colors are equal to the corresponding standard color proportions, proceed to the subsequent steps.

5. The cloud-based design collaboration platform system based on customized model training according to claim 4 is characterized in that: The analysis process of the color analysis module also includes: Calculate the red first-order color moment and the red second-order color moment of the corresponding color in the first clothing design area; The first-order color moment of red is: Among them, o is the color number, i represents the number of the pixel in the first clothing design area, Rio represents the R value component of the RGB value corresponding to the pixel, and Uro represents the color deviation of the corresponding clothing design area. N is the upper limit value of i, and N is an Arabic numeral; The second-order color moment is: The red first-order color moment is used to characterize the average response intensity of the red component in different colors in the first clothing design area, and the red second-order color moment is used to represent the response variance of the red component in different colors in the first clothing design area; The red excellence values ​​RYXo of different colors in the first clothing design area are calculated by the formula, and the specific formula is as follows: RYXo=1 / (A1×YJro+A2×EJro); where A1 and A2 are weight coefficients with fixed values, and A2>A1; Similarly, calculate the green excellence value and blue excellence value of different colors in the first clothing design area; The red excellent value, green excellent value and blue excellent value of different colors in the first clothing design area are calculated by the formula to obtain the color excellent value of all colors in the first clothing design area. The formula is: Color excellent value = red excellent value * weight coefficient 1 + green excellent value * weight coefficient 2 + blue excellent value * weight coefficient 3; Traverse and compare the color excellence values ​​of all colors in the first clothing design area, obtain the minimum value among the color excellence values, and use the minimum value as the first color excellence value of the first clothing design area; Similarly, the second to seventh color excellence values ​​corresponding to the second to seventh clothing design areas are obtained.

6. The cloud-based design collaboration platform system based on customized model training according to claim 5, characterized in that: The analysis process of the texture analysis module is as follows: Obtain the RGB value of the pixel point in the first clothing design area, and add the R value component, the G value component, and the B value component in the RGB value of the same pixel point to obtain the sum of the R value component, the G value component, and the B value component as the pixel value XSZi of the corresponding pixel point; Take every nine pixels as an independent window, record the surrounding eight pixels as edge pixels, and record the middle pixel as the center pixel; start with the edge pixel in the upper left corner, rotate clockwise, and compare the pixel values ​​of the edge pixels with the pixel values ​​of the center pixel; If the pixel value of the edge pixel is greater than or equal to the pixel value of the center pixel, the gray value of the edge pixel is set to 1; if the pixel value of the edge pixel is less than the pixel value of the center pixel, the gray value of the edge pixel is set to 0; The pixel points with a gray value of 1 are connected to obtain a preliminary texture image corresponding to the first clothing design area.

7. The cloud-based design collaboration platform system based on customized model training according to claim 6, characterized in that: The analysis process of the texture analysis module also includes: Calculate the Hamming distance between the prepared texture image and multiple texture images preset in the database by using an image hash algorithm; If the Hamming distance of the first clothing design area is greater than or equal to the first Hamming threshold, an abnormal signal is generated; If the Hamming distance of the first clothing design region is less than the first Hamming threshold and greater than or equal to the second Hamming threshold, the texture similarity level of the first clothing design region is recorded as first-level similarity; If the Hamming distance of the first clothing design area is less than the second Hamming threshold, the texture similarity level of the first clothing design area is recorded as secondary similarity; wherein the first Hamming threshold is greater than the second Hamming threshold, and the texture similarity corresponding to the primary similarity is less than the texture similarity corresponding to the secondary similarity; Similarly, the texture similarity levels of the second clothing design area to the seventh clothing design area are obtained.

8. The cloud-based design collaboration platform system based on customized model training according to claim 7, characterized in that: The analysis process of the comprehensive analysis module is as follows: Obtain the first color excellent value SC1 to the seventh color excellent value SC7 corresponding to the clothing design drawing; Obtaining texture similarity levels of the first clothing design region to the seventh clothing design region; Calculate the clothing excellence value of the first clothing design area, FY1 = SC1 × λ, where λ is the texture similarity weight value, and the texture similarity weight value corresponding to the first-level similarity is less than the texture similarity weight value corresponding to the second-level similarity; Similarly, the clothing excellence values ​​of the second clothing design area to the seventh clothing design area are calculated; Compare the clothing excellence values ​​of the first to seventh clothing design areas with the clothing excellence threshold; if the clothing excellence values ​​of all clothing design areas are greater than the clothing excellence threshold, automatically generate a physical clothing image corresponding to the clothing design image; If the clothing excellence value of any of the first to seventh clothing design areas is less than the clothing excellence threshold, a modification signal for the corresponding clothing design area is generated.

Images