CN103886115A - Building and calling method of three-dimension virtual body form based on different body types - Google Patents
Building and calling method of three-dimension virtual body form based on different body types Download PDFInfo
- Publication number
- CN103886115A CN103886115A CN201210556599.8A CN201210556599A CN103886115A CN 103886115 A CN103886115 A CN 103886115A CN 201210556599 A CN201210556599 A CN 201210556599A CN 103886115 A CN103886115 A CN 103886115A
- Authority
- CN
- China
- Prior art keywords
- chest
- angle
- buttocks
- central point
- people
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000013528 artificial neural network Methods 0.000 claims abstract description 33
- 238000012935 Averaging Methods 0.000 claims abstract description 4
- 210000001217 buttock Anatomy 0.000 claims description 76
- 210000001015 abdomen Anatomy 0.000 claims description 55
- 238000005259 measurement Methods 0.000 claims description 39
- 230000000877 morphologic effect Effects 0.000 claims description 17
- 230000000392 somatic effect Effects 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 7
- 210000003050 axon Anatomy 0.000 claims description 6
- 210000003127 knee Anatomy 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 3
- 238000003708 edge detection Methods 0.000 claims description 2
- 238000013461 design Methods 0.000 abstract description 8
- 238000004519 manufacturing process Methods 0.000 abstract description 5
- 210000000038 chest Anatomy 0.000 description 124
- 239000000543 intermediate Substances 0.000 description 17
- 241000288105 Grus Species 0.000 description 3
- 210000002414 leg Anatomy 0.000 description 3
- 230000008929 regeneration Effects 0.000 description 2
- 238000011069 regeneration method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a building and calling method of a three-dimension virtual body form based on different body types. The method includes the steps of firstly, categorizing acquired body types by using height/chest circumference, lateral forms, and chest and waist difference as indexes, and using an averaging method to obtain the body data of a middle body of each body type; secondly, building a BP neural network to convert measured body data into body type modeling parameters; thirdly, compiling the body type interface file of the middle body of each body type to virtually reproduce the middle body of each body type; fourthly, respectively building judging formulas of the indexes of height/chest circumference, lateral forms, and chest and waist difference through counting and analysis, and calling corresponding virtual body form through formula judging. By the method, virtual body form reproduction and calling of corresponding body form can be achieved, fitting design of garment can be achieved by the virtual body forms, a novel method is provided for garment MTM production, and the method is widely application to garment customizing platforms.
Description
Technical field
The present invention relates to a kind of foundation and call method thereof of the three-dimensional virtual human platform based on different builds, belong to clothes industry digitizing technique field.
Background technology
Along with the raising of people's living standard, the clothes MTM mode of production based on clothes personalized design is arisen at the historic moment.In the personalized design of clothes, the zoariumization design of the garment dimension based on client's build is a very important aspect.Therefore,, as emerging garment fit evaluation means, based on the three-dimensional virtual fitting of customization virtual human body and customization virtual costume, received increasing concern.How, by several human dimensions simple and easy to get, generate the dummy body form of respective client build, thereby be the large study hotspot in this area one by the Dam Configuration Design that closes that this people's platform carries out clothes.
Summary of the invention
The object of this invention is to provide a kind of foundation and call method thereof of the three-dimensional virtual human platform based on different builds, to realize the Dam Configuration Design that closes that carries out clothes by dummy body form, provide a kind of new method for meeting clothes MTM production.
For achieving the above object, the technical solution used in the present invention is as follows:
Foundation and the call method thereof of three-dimensional virtual human platform based on different builds, comprise the steps:
The first step, the buman body type obtaining is classified as index so that height/chest measurement, side form, chest waist are poor, and use the method for averaging to obtain the somatic data of each build intermediate;
The Changing Pattern of the virtual human body in second step, analyzing three-dimensional fitting software, take physical characteristic parameter as input parameter, take the shape parameter that is applied to three-dimensional modeling as output parameter, build BP neural network, realize the conversion of anthropometric data to build modeling parameters;
The 3rd step, analyze the coded system of virtual human body interface document in this three-dimensional fitting software, write the build interface document of each build intermediate, realize the virtual reappearance of each build intermediate;
The 4th step, by statistical study, set up respectively the discrimination formula of height/chest measurement index, people's body side surface morphological index and the poor index of chest waist; Differentiate by formula, realize calling respective virtual people platform.
As a kind of preferred version, described side morphological index refers to that people's body side surface abdomen salient point is to the line of chest salient point and the angle of surface level and people's body side surface stern salient point to the back of the body line of salient point and the angle of surface level.
As further preferred version, the acquisition of described side morphological index is under the environment of Matlab(matrix experiment chamber (matrixlaboratory), use Roberts edge detection operator (one of operator), find the edge of people's body side surface photo, identify matrix of edge salient point by matrix operation, by salient point coordinate, obtain described side morphological index.
As preferred version further, before identification matrix of edge salient point, first image is divided into two by horizontal 1:1, be divided into two by longitudinal 3:4.
As a kind of preferred version, the operation of the first step comprises:
A) use Scanworx spatial digitizer (one of high precision scanner), the anthropometric data that automatic acquisition is whole, and increase following 36 measure the items of human body by the calculating of manual measurement and derived variable: front fillet, rear fillet, average fillet, side neck is put the distance/front length of chest point, BP dot spacing/chest measurement, abdomen coign, Attacking Midfielder's central point is to the angle of chest both sides, side waist central point is to the angle of chest both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of chest both sides to chest both sides, Attacking Midfielder is to the vertical angle of chest, Attacking Midfielder's central point is to the angle of buttocks both sides, side waist central point is to the angle of buttocks both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of buttocks both sides to buttocks both sides, Attacking Midfielder's central point is to the angle of angle/Attacking Midfielder central point chest both sides of buttocks both sides, the angle of side waist central point to angle/side waist central point of buttocks both sides to chest both sides, low back is to the vertical angle of stern, back of the body length/height, side knee central point is to the vertical angle of waist central point, side waist central point is to the angle of stern central point, lower body shaft angle, axon angle above the waist, before upper body, angle, center and low back are to collare angle,
B) take height/chest measurement as index, human body is divided into height, in, short three classes, wherein high is that relative chest measurement is high, in be in relative chest measurement, short is that relative chest measurement is short; With people's body side surface morphological index: abdomen bump to the line of chest salient point with the angle of surface level, stern salient point to carrying on the back the line of salient point and the angle of surface level, human body is divided into AC, AB, AA, BC, BB, BA, CC, CB, CA nine classes; Wherein: AC refers to that the buttocks of people's body side surface protrudes than belly than back protrusion, chest; AB refers to that the buttocks of people's body side surface is similar on a vertical vertical line than back protrusion, chest and belly; AA refers to that the buttocks of people's body side surface protrudes than chest than back protrusion, belly; BC refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly; BB refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest and belly are similar on a vertical line; BA refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly; The back that CC refers to people's body side surface than buttocks protrude, chest protrudes than belly; CB refers to that the back of people's body side surface is similar on a vertical vertical line than buttocks protrusion, chest and belly; The back that CA refers to people's body side surface than buttocks protrude, belly protrudes than chest; In addition, human body is divided into Y, A, B, C tetra-classes so that chest waist is poor as index by national standard; Obtain altogether 108 class builds;
C) somatic data of 108 class builds is carried out to statistical study, ask after its mean value, obtain the somatic data of each build intermediate.
As a kind of preferred version, the operation of second step comprises:
D) the some shape adjustment parameters of virtual human body of controlling in three-dimensional fitting software change, and while observing this parameter variation, the Changing Pattern of virtual human body build, obtains the larger physical characteristic parameter of body surface impact;
E) select the input parameter of 36 idiotype characteristic parameters as BP neural network, selecting 11 shape parameters that are applied to three-dimensional modeling is output parameter, builds the initial model of BP network under Matlab environment;
F) pass through network training, analyze minimax method, variance method for normalizing, and three kinds of methods of non-processor are carried out the impact on neural network forecast ability after standardization to data respectively, analyze the different network numbers of plies impact on neural network forecast ability is set, analyze the impact of different training function setup on neural network forecast ability, analyze different node transport functions the impact on neural network forecast ability is set, analyze different the number of hidden nodes the impact on neural network forecast ability is set, analyze heterogeneous networks input and output parameter the impact on neural network forecast ability is set, obtain best network structure, set up physical characteristic parameter and the shape adjustment nonlinearity in parameters mapping relations for three-dimensional fitting software.
As further preferred version, 36 described idiotype characteristic parameters are: front fillet, rear fillet, average fillet, side neck is put the distance of chest point, side neck is put the distance/front length of chest point, BP dot spacing, BP dot spacing/chest measurement, abdomen coign, Attacking Midfielder's central point is to the angle of chest both sides, side waist central point is to the angle of chest both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of chest both sides to chest both sides, Attacking Midfielder is to the vertical angle of chest, Attacking Midfielder's central point is to the angle of buttocks both sides, side waist central point is to the angle of buttocks both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of buttocks both sides to buttocks both sides, Attacking Midfielder's central point is to the angle of angle/Attacking Midfielder central point chest both sides of buttocks both sides, the angle of side waist central point to angle/side waist central point of buttocks both sides to chest both sides, low back is to the vertical angle of stern, the back of the body is long, back of the body length/height, side knee central point is to the vertical angle of waist central point, side waist central point is to the angle of stern central point, lower body shaft angle, axon angle above the waist, angle, center before upper body, low back is to collare angle, highly, neck circumference, shoulder breadth, chest measurement, under chest, enclose, waistline, lower hip circumference, upper hip circumference, outer leg length and interior length.
As further preferred version, 11 described shape parameters that are applied to three-dimensional modeling are: shoulder parameter, chest locations parameter, chest peak distance parameter, abdomen parameter, health width parameter, buttocks width parameter, upper part of the body long parameter, pose parameter, lower part of the body pose parameter, bottom position parameter and upper body pose parameter.
As a kind of preferred version, the operation of the 3rd step comprises:
G) coded system of the interface document of human parameters modeling in analyzing three-dimensional fitting software;
H) set up third party software platform, built-in neural network structure, by by corresponding physical characteristic data file input, through the parameter conversion of neural network, generates the people's platform interface document that is applied to various build intermediates in three-dimensional fitting software automatically.
As a kind of preferred version, the operation of the 4th step comprises:
I) under Matlab environment, set up the automatic identification algorithm of people's body side surface morphological index: side colour picture is converted to gray-scale map, again by Robort operator, find the edge of people's body side surface photo, the part that has edge lines that reads human body edge image by Matlab is 1, and all the other are 0; Image is divided into 4 parts, and wherein horizontal ratio is 1:1, and longitudinal ration of division is 3:4; Utilize the determinant of digital picture, find first method that occurs 1 ranks and find the salient point of every image; Salient point position after merging by determinant, obtains the coordinate figure of each salient point; Calculate people's body side surface abdomen salient point by coordinate figure and arrive back of the body salient point line and horizontal angle to chest salient point line and horizontal angle, people's body side surface stern salient point;
The human body of j) height/chest measurement, side morphological index, the poor index of chest waist being classified, the statistics of point anthropoid boundary value that carries out respectively obtaining under each sorting technique, and set up corresponding build discrimination formula; By build discrimination formula, build is belonged to class and differentiate, call corresponding interface document, generate corresponding intermediate people platform.
Compared with prior art, the present invention has following conspicuousness progress and beneficial effect:
The present invention carries out the classification of K-Mean average by the somatic data that three-dimensional human body measurement is obtained, and the data of adding up the each position of corresponding human body intermediate value; By building characteristics of human body parameter to the coded system of four layers of BP model of shape adjustment parameter and the interface document of virtual human body that is applied to existing three-dimensional fitting software, realize the regeneration of dummy body form; By setting up the fuzzy discrimination formula of each human body classification indicators, realize buman body type and differentiate, realize respective virtual people tableland has been called; The present invention can carry out by dummy body form the Dam Configuration Design that closes of clothes, provides a kind of new method for meeting clothes MTM production, can be widely used in custom made clothing platform.
Embodiment
Below in conjunction with accompanying drawing, the present invention is illustrated in further detail:
Three-dimensional fitting software described in the present embodiment is the V-stitcher software of Gerber company of the U.S., and described Matlab is Matlab7.1 version.
Foundation and the call method thereof of a kind of three-dimensional virtual human platform based on different builds provided by the invention, comprise the steps:
The first step, the buman body type obtaining is classified as index so that height/chest measurement, side form, chest waist are poor, and use the method for averaging to obtain the somatic data of each build intermediate;
The Changing Pattern of the virtual human body in second step, analyzing three-dimensional fitting software, take physical characteristic parameter as input parameter, take the shape parameter that is applied to three-dimensional modeling as output parameter, build BP neural network, realize the conversion of anthropometric data to build modeling parameters;
The 3rd step, analyze the coded system of virtual human body interface document in this three-dimensional fitting software, write the build interface document of each build intermediate, realize the virtual reappearance of each build intermediate;
The 4th step, by statistical study, set up respectively the discrimination formula of height/chest measurement index, people's body side surface morphological index and the poor index of chest waist; Differentiate by formula, realize calling respective virtual people platform.
As a kind of preferred version, described method comprises following operation:
The first step:
A) use Scanworx spatial digitizer (one of high precision scanner), the anthropometric data that automatic acquisition is whole, and increase following 36 measure the items of human body by the calculating of manual measurement and derived variable: front fillet, rear fillet, average fillet, side neck is put the distance/front length of chest point, BP dot spacing/chest measurement, abdomen coign, Attacking Midfielder's central point is to the angle of chest both sides, side waist central point is to the angle of chest both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of chest both sides to chest both sides, Attacking Midfielder is to the vertical angle of chest, Attacking Midfielder's central point is to the angle of buttocks both sides, side waist central point is to the angle of buttocks both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of buttocks both sides to buttocks both sides, Attacking Midfielder's central point is to the angle of angle/Attacking Midfielder central point chest both sides of buttocks both sides, the angle of side waist central point to angle/side waist central point of buttocks both sides to chest both sides, low back is to the vertical angle of stern, back of the body length/height, side knee central point is to the vertical angle of waist central point, side waist central point is to the angle of stern central point, lower body shaft angle, axon angle above the waist, before upper body, angle, center and low back are to collare angle,
B) take height/chest measurement as index, human body is divided into height, in, short three classes, wherein high is that relative chest measurement is high, in be in relative chest measurement, short is that relative chest measurement is short; With people's body side surface morphological index: abdomen bump to the line of chest salient point with the angle of surface level, stern salient point to carrying on the back the line of salient point and the angle of surface level, human body is divided into AC, AB, AA, BC, BB, BA, CC, CB, CA nine classes; Wherein: AC refers to that the buttocks of people's body side surface protrudes than belly than back protrusion, chest; AB refers to that the buttocks of people's body side surface is similar on a vertical vertical line than back protrusion, chest and belly; AA refers to that the buttocks of people's body side surface protrudes than chest than back protrusion, belly; BC refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly; BB refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest and belly are similar on a vertical line; BA refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly; The back that CC refers to people's body side surface than buttocks protrude, chest protrudes than belly; CB refers to that the back of people's body side surface is similar on a vertical vertical line than buttocks protrusion, chest and belly; The back that CA refers to people's body side surface than buttocks protrude, belly protrudes than chest; In addition, human body is divided into Y, A, B, C tetra-classes so that chest waist is poor as index by national standard; Obtain altogether 108 class builds;
C) somatic data of 108 class builds is carried out to statistical study, ask after its mean value, obtain the somatic data of each build intermediate.
Second step:
D) the some shape adjustment parameters of virtual human body of controlling in three-dimensional fitting software change, and while observing this parameter variation, the Changing Pattern of virtual human body build, obtains the larger physical characteristic parameter of body surface impact;
E) select the input parameter of 36 idiotype characteristic parameters as BP neural network, selecting 11 shape parameters that are applied to three-dimensional modeling is output parameter, builds the initial model of BP network under Matlab environment;
F) pass through network training, analyze minimax method, variance method for normalizing, and three kinds of methods of non-processor are carried out the impact on neural network forecast ability after standardization to data respectively, analyze the different network numbers of plies impact on neural network forecast ability is set, analyze the impact of different training function setup on neural network forecast ability, analyze different node transport functions the impact on neural network forecast ability is set, analyze different the number of hidden nodes the impact on neural network forecast ability is set, analyze heterogeneous networks input and output parameter the impact on neural network forecast ability is set, obtain best network structure, set up physical characteristic parameter and the shape adjustment nonlinearity in parameters mapping relations for three-dimensional fitting software,
36 described idiotype characteristic parameters are: front fillet, rear fillet, average fillet, side neck is put the distance of chest point, side neck is put the distance/front length of chest point, BP dot spacing, BP dot spacing/chest measurement, abdomen coign, Attacking Midfielder's central point is to the angle of chest both sides, side waist central point is to the angle of chest both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of chest both sides to chest both sides, Attacking Midfielder is to the vertical angle of chest, Attacking Midfielder's central point is to the angle of buttocks both sides, side waist central point is to the angle of buttocks both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of buttocks both sides to buttocks both sides, Attacking Midfielder's central point is to the angle of angle/Attacking Midfielder central point chest both sides of buttocks both sides, the angle of side waist central point to angle/side waist central point of buttocks both sides to chest both sides, low back is to the vertical angle of stern, the back of the body is long, back of the body length/height, side knee central point is to the vertical angle of waist central point, side waist central point is to the angle of stern central point, lower body shaft angle, axon angle above the waist, angle, center before upper body, low back is to collare angle, highly, neck circumference, shoulder breadth, chest measurement, under chest, enclose, waistline, lower hip circumference, upper hip circumference, outer leg length and interior length,
11 described shape parameters that are applied to three-dimensional modeling are: shoulder parameter, chest locations parameter, chest peak distance parameter, abdomen parameter, health width parameter, buttocks width parameter, upper part of the body long parameter, pose parameter, lower part of the body pose parameter, bottom position parameter and upper body pose parameter.
The 3rd step:
G) coded system of the interface document of human parameters modeling in analyzing three-dimensional fitting software;
H) set up third party software platform, built-in neural network structure, by by corresponding physical characteristic data file input, through the parameter conversion of neural network, generates the people's platform interface document that is applied to various build intermediates in three-dimensional fitting software automatically.
The 4th step:
I) under Matlab environment, set up the automatic identification algorithm of people's body side surface morphological index: side colour picture is converted to gray-scale map, again by Robort operator, find the edge of people's body side surface photo, the part that has edge lines that reads human body edge image by Matlab is 1, and all the other are 0; Image is divided into 4 parts, and wherein horizontal ratio is 1:1, and longitudinal ration of division is 3:4; Utilize the determinant of digital picture, find first method that occurs 1 ranks and find the salient point of every image; Salient point position after merging by determinant, obtains the coordinate figure of each salient point; Calculate people's body side surface abdomen salient point by coordinate figure and arrive back of the body salient point line and horizontal angle to chest salient point line and horizontal angle, people's body side surface stern salient point;
The human body of j) height/chest measurement, side morphological index, the poor index of chest waist being classified, the statistics of point anthropoid boundary value that carries out respectively obtaining under each sorting technique, and set up corresponding build discrimination formula; By build discrimination formula, build is belonged to class and differentiate, call corresponding interface document, generate corresponding intermediate people platform.
Embodiment
Take height/chest measurement as index, human body is divided into height, in, short three classes, wherein high is that relative chest measurement is high, in be in relative chest measurement, short is that relative chest measurement is short, with people's body side surface index: abdomen bump to the line of chest salient point with the angle of surface level, stern salient point to carrying on the back the line of salient point and the angle of surface level, human body is divided into AC, AB, AA, BC, BB, BA, CC, CB, CA nine classes, wherein: AC refers to that the buttocks of people's body side surface protrudes than back, chest protrudes than belly, AB refers to that the buttocks of people's body side surface protrudes than back, chest and belly are similar on a vertical vertical line, AA refers to that the buttocks of people's body side surface protrudes than back, belly protrudes than chest, BC refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly, BB refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest and belly are similar on a vertical line, BA refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly, CC refers to that the back of people's body side surface protrudes than buttocks, chest protrudes than belly, CB refers to that the back of people's body side surface protrudes than buttocks, chest and belly are similar on a vertical vertical line, CA refers to that the back of people's body side surface protrudes than buttocks, belly protrudes than chest.So that chest waist is poor, human body is divided into Y, A, B, C tetra-classes by national standard.Obtain altogether 108 class builds.
Build data to 108 class builds are carried out statistical study, ask after its mean value, obtain corresponding intermediate data.
With shoulder parameter, chest locations parameter, chest peak distance parameter, abdomen parameter, health width parameter, buttocks width parameter, long parameter above the waist, pose parameter, lower part of the body pose parameter, bottom position parameter, upper body pose parameter is the output parameter of 11 neural networks, with 36 idiotype characteristic parameters: front fillet, rear fillet, average fillet, side neck is put the distance of chest point, side neck is put the distance/front length of chest point, BP dot spacing, BP dot spacing/chest measurement, abdomen coign, Attacking Midfielder's central point is to the angle of chest both sides, side waist central point is to the angle of chest both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of chest both sides to chest both sides, Attacking Midfielder is to the vertical angle of chest, Attacking Midfielder's central point is to the angle of buttocks both sides, side waist central point is to the angle of buttocks both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of buttocks both sides to buttocks both sides, Attacking Midfielder's central point is to the angle of angle/Attacking Midfielder central point chest both sides of buttocks both sides, the angle of side waist central point to angle/side waist central point of buttocks both sides to chest both sides, low back is to the vertical angle of stern, the back of the body is long, back of the body length/height, side knee central point is to the vertical angle of waist central point, side waist central point is to the angle of stern central point, lower body shaft angle, axon angle above the waist, angle, center before upper body, low back is to collare angle, highly, neck circumference, shoulder breadth, chest measurement, under chest, enclose, waistline, lower hip circumference, upper hip circumference, outer leg is long, interior length is input parameter, under the environment of Matlab7.1, build four layers of BP neural network, take minimax method to be normalized input data, the training function setup of network be trainlm}, the node transport function of network is set to { logsig, purelin, logsig}, epoch is set to 1000, and learning rate is set to 0.01, realizes the conversion of the shape adjustment parameter of fitting software from characteristics of human body's parameter to applying three-dimensional.
The coded system of the interface document of human parameters modeling in analyzing three-dimensional fitting software, set up third party software platform, built-in neural network structure, by corresponding physical characteristic data file is inputted, through the parameter conversion of neural network, automatically generate the people's platform interface document that is applied to various build intermediates in three-dimensional fitting software.
Under the environment of Matlab7.1, set up the automatic identification algorithm of people's body side surface index: first side colour picture is converted to gray-scale map, again by Robort operator, find the edge of people's body side surface photo, the part that has edge lines that reads human body edge image by Matlab is like this 1, and all the other are 0.In order to calculate fast the chest salient point, back salient point, buttocks salient point and the belly salient point that obtain human body, image is divided into 4 parts, wherein grid scale is 1:1, longitudinal ration of division is 3:4.Utilize the determinant of digital picture, find first method that occurs 1 ranks and find the salient point of every image.Salient point position after merging by determinant, obtains the coordinate figure of each salient point.Calculate by coordinate figure: people's body side surface abdomen salient point arrives back of the body salient point line and horizontal angle to chest salient point line and horizontal angle, people's body side surface stern salient point.
The human body of then height/chest measurement, side morphological index, the poor index of chest waist being classified, the statistics of point anthropoid boundary value that carries out respectively obtaining under each sorting technique, and set up corresponding build discrimination formula.By build discrimination formula, build is belonged to class and differentiate, call corresponding interface document, generate corresponding intermediate people platform.
Visible in sum: the present invention carries out the classification of K-Mean average by the somatic data that three-dimensional human body measurement is obtained, and the data of adding up the each position of corresponding human body intermediate value; By building characteristics of human body parameter to the coded system of four layers of BP model of shape adjustment parameter and the interface document of virtual human body that is applied to existing three-dimensional fitting software, realize the regeneration of dummy body form; By setting up the fuzzy discrimination formula of each human body classification indicators, realize buman body type and differentiate, realize respective virtual people tableland has been called; The present invention can carry out by dummy body form the Dam Configuration Design that closes of clothes, provides a kind of new method for meeting clothes MTM production, can be widely used in custom made clothing platform.
Finally be necessary to be pointed out that at this; above-mentioned explanation is only for being described in further detail technical scheme of the present invention; can not be interpreted as limiting the scope of the invention, some nonessential improvement that those skilled in the art's foregoing according to the present invention is made and adjustment all belong to protection scope of the present invention.
Claims (7)
1. foundation and the call method thereof of the three-dimensional virtual human platform based on different builds, is characterized in that, comprises the steps:
The first step, the buman body type obtaining is classified as index so that height/chest measurement, side form, chest waist are poor, and use the method for averaging to obtain the somatic data of each build intermediate;
The Changing Pattern of the virtual human body in second step, analyzing three-dimensional fitting software, take physical characteristic parameter as input parameter, take the shape parameter that is applied to three-dimensional modeling as output parameter, build BP neural network, realize the conversion of anthropometric data to build modeling parameters;
The 3rd step, analyze the coded system of virtual human body interface document in this three-dimensional fitting software, write the build interface document of each build intermediate, realize the virtual reappearance of each build intermediate;
The 4th step, by statistical study, set up respectively the discrimination formula of height/chest measurement index, people's body side surface morphological index and the poor index of chest waist; Differentiate by formula, realize calling respective virtual people platform.
2. foundation and the call method thereof of the three-dimensional virtual human platform based on different builds according to claim 1, is characterized in that: described side morphological index refers to that people's body side surface abdomen salient point is to the line of chest salient point and the angle of surface level and people's body side surface stern salient point to the back of the body line of salient point and the angle of surface level.
3. foundation and the call method thereof of the three-dimensional virtual human platform based on different builds according to claim 2, it is characterized in that: the acquisition of described side morphological index is under Matlab environment, use Roberts edge detection operator, find the edge of people's body side surface photo, identify matrix of edge salient point by matrix operation, by salient point coordinate, obtain described side morphological index.
4. foundation and the call method thereof of the three-dimensional virtual human platform based on different builds according to claim 3, is characterized in that: before identification matrix of edge salient point, first image is divided into two by horizontal 1:1, is divided into two by longitudinal 3:4.
5. foundation and the call method thereof of the three-dimensional virtual human platform based on different builds according to claim 1, is characterized in that, the operation of the first step comprises:
A) use Scanworx spatial digitizer, the anthropometric data that automatic acquisition is whole, and increase following 36 measure the items of human body by the calculating of manual measurement and derived variable: front fillet, rear fillet, average fillet, side neck is put the distance/front length of chest point, BP dot spacing/chest measurement, abdomen coign, Attacking Midfielder's central point is to the angle of chest both sides, side waist central point is to the angle of chest both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of chest both sides to chest both sides, Attacking Midfielder is to the vertical angle of chest, Attacking Midfielder's central point is to the angle of buttocks both sides, side waist central point is to the angle of buttocks both sides, the angle of Attacking Midfielder's central point to angle/side waist central point of buttocks both sides to buttocks both sides, Attacking Midfielder's central point is to the angle of angle/Attacking Midfielder central point chest both sides of buttocks both sides, the angle of side waist central point to angle/side waist central point of buttocks both sides to chest both sides, low back is to the vertical angle of stern, back of the body length/height, side knee central point is to the vertical angle of waist central point, side waist central point is to the angle of stern central point, lower body shaft angle, axon angle above the waist, before upper body, angle, center and low back are to collare angle,
B) take height/chest measurement as index, human body is divided into height, in, short three classes, wherein high is that relative chest measurement is high, in be in relative chest measurement, short is that relative chest measurement is short; With people's body side surface morphological index: abdomen bump to the line of chest salient point with the angle of surface level, stern salient point to carrying on the back the line of salient point and the angle of surface level, human body is divided into AC, AB, AA, BC, BB, BA, CC, CB, CA nine classes; Wherein: AC refers to that the buttocks of people's body side surface protrudes than belly than back protrusion, chest; AB refers to that the buttocks of people's body side surface is similar on a vertical vertical line than back protrusion, chest and belly; AA refers to that the buttocks of people's body side surface protrudes than chest than back protrusion, belly; BC refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly; BB refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest and belly are similar on a vertical line; BA refers to that the buttocks of people's body side surface and back are similar on a vertical vertical line, chest protrudes than belly; The back that CC refers to people's body side surface than buttocks protrude, chest protrudes than belly; CB refers to that the back of people's body side surface is similar on a vertical vertical line than buttocks protrusion, chest and belly; The back that CA refers to people's body side surface than buttocks protrude, belly protrudes than chest; In addition, human body is divided into Y, A, B, C tetra-classes so that chest waist is poor as index by national standard; Obtain altogether 108 class builds;
C) somatic data of 108 class builds is carried out to statistical study, ask after its mean value, obtain the somatic data of each build intermediate.
6. foundation and the call method thereof of the three-dimensional virtual human platform based on different builds according to claim 1, is characterized in that, the operation of second step comprises:
D) the some shape adjustment parameters of virtual human body of controlling in three-dimensional fitting software change, and while observing this parameter variation, the Changing Pattern of virtual human body build, obtains the larger physical characteristic parameter of body surface impact;
E) select the input parameter of 36 idiotype characteristic parameters as BP neural network, selecting 11 shape parameters that are applied to three-dimensional modeling is output parameter, builds the initial model of BP network under Matlab environment;
F) pass through network training, analyze minimax method, variance method for normalizing, and three kinds of methods of non-processor are carried out the impact on neural network forecast ability after standardization to data respectively, analyze the different network numbers of plies impact on neural network forecast ability is set, analyze the impact of different training function setup on neural network forecast ability, analyze different node transport functions the impact on neural network forecast ability is set, analyze different the number of hidden nodes the impact on neural network forecast ability is set, analyze heterogeneous networks input and output parameter the impact on neural network forecast ability is set, obtain best network structure, set up physical characteristic parameter and the shape adjustment nonlinearity in parameters mapping relations for three-dimensional fitting software.
7. foundation and the call method thereof of the three-dimensional virtual human platform based on different builds according to claim 6, is characterized in that, 36 described idiotype characteristic parameters are: front fillet, rear fillet, average fillet, side neck are put the distance of chest point.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210556599.8A CN103886115B (en) | 2012-12-20 | 2012-12-20 | The foundation of a kind of three-dimensional virtual human platform based on different building shape and call method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210556599.8A CN103886115B (en) | 2012-12-20 | 2012-12-20 | The foundation of a kind of three-dimensional virtual human platform based on different building shape and call method thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103886115A true CN103886115A (en) | 2014-06-25 |
CN103886115B CN103886115B (en) | 2016-12-28 |
Family
ID=50955007
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210556599.8A Expired - Fee Related CN103886115B (en) | 2012-12-20 | 2012-12-20 | The foundation of a kind of three-dimensional virtual human platform based on different building shape and call method thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103886115B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104881892A (en) * | 2015-06-04 | 2015-09-02 | 厦门唯尔酷信息技术有限公司 | Model deformation method and model deformation system |
CN106418848A (en) * | 2016-08-04 | 2017-02-22 | 上海工程技术大学 | Body type classification method for customizing shirts |
CN107895145A (en) * | 2017-10-31 | 2018-04-10 | 南京信息工程大学 | Method based on convolutional neural networks combination super-Gaussian denoising estimation finger stress |
CN108009577A (en) * | 2017-11-29 | 2018-05-08 | 南京工业大学 | A kind of implementation method of virtual mirror |
CN108985657A (en) * | 2018-08-06 | 2018-12-11 | 百度在线网络技术(北京)有限公司 | Evaluation method, device and the server of personage's clothing collocation |
CN109636907A (en) * | 2018-12-13 | 2019-04-16 | 谷东科技有限公司 | A kind of terrain reconstruction method and system based on AR glasses |
CN110135078A (en) * | 2019-05-17 | 2019-08-16 | 上海凌笛数码科技有限公司 | A kind of human parameters automatic generation method based on machine learning |
CN110533767A (en) * | 2019-08-15 | 2019-12-03 | 东华大学 | Bulging body men's shirt template parameter optimization method based on three-dimensional garment design |
CN111783182A (en) * | 2020-07-07 | 2020-10-16 | 恒信东方文化股份有限公司 | Modeling method and system of three-dimensional virtual mannequin |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120086783A1 (en) * | 2010-06-08 | 2012-04-12 | Raj Sareen | System and method for body scanning and avatar creation |
-
2012
- 2012-12-20 CN CN201210556599.8A patent/CN103886115B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120086783A1 (en) * | 2010-06-08 | 2012-04-12 | Raj Sareen | System and method for body scanning and avatar creation |
Non-Patent Citations (2)
Title |
---|
郑爱花: "<基于BP神经网络的服装号型推荐方法研究>", 《中国优秀硕士学位论文全文数据库 工程科技I辑》 * |
黄英: "N我国服装人台研究现状概述", 《浙江纺织服装职业技术学院学报》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104881892A (en) * | 2015-06-04 | 2015-09-02 | 厦门唯尔酷信息技术有限公司 | Model deformation method and model deformation system |
CN104881892B (en) * | 2015-06-04 | 2017-10-27 | 厦门唯尔酷信息技术有限公司 | Model deformation method and its system |
CN106418848A (en) * | 2016-08-04 | 2017-02-22 | 上海工程技术大学 | Body type classification method for customizing shirts |
CN107895145A (en) * | 2017-10-31 | 2018-04-10 | 南京信息工程大学 | Method based on convolutional neural networks combination super-Gaussian denoising estimation finger stress |
CN108009577A (en) * | 2017-11-29 | 2018-05-08 | 南京工业大学 | A kind of implementation method of virtual mirror |
CN108985657A (en) * | 2018-08-06 | 2018-12-11 | 百度在线网络技术(北京)有限公司 | Evaluation method, device and the server of personage's clothing collocation |
CN109636907A (en) * | 2018-12-13 | 2019-04-16 | 谷东科技有限公司 | A kind of terrain reconstruction method and system based on AR glasses |
CN110135078A (en) * | 2019-05-17 | 2019-08-16 | 上海凌笛数码科技有限公司 | A kind of human parameters automatic generation method based on machine learning |
WO2020232917A1 (en) * | 2019-05-17 | 2020-11-26 | 上海凌笛数码科技有限公司 | Automatic human body parameter generation method based on machine learning |
CN110135078B (en) * | 2019-05-17 | 2023-03-14 | 浙江凌迪数字科技有限公司 | Human body parameter automatic generation method based on machine learning |
CN110533767A (en) * | 2019-08-15 | 2019-12-03 | 东华大学 | Bulging body men's shirt template parameter optimization method based on three-dimensional garment design |
CN111783182A (en) * | 2020-07-07 | 2020-10-16 | 恒信东方文化股份有限公司 | Modeling method and system of three-dimensional virtual mannequin |
CN111783182B (en) * | 2020-07-07 | 2024-03-01 | 恒信东方文化股份有限公司 | Modeling method and system of three-dimensional virtual mannequin |
Also Published As
Publication number | Publication date |
---|---|
CN103886115B (en) | 2016-12-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103886115A (en) | Building and calling method of three-dimension virtual body form based on different body types | |
CN106250874B (en) | Recognition methods and the device of a kind of dress ornament and carry-on articles | |
CN103186904B (en) | Picture contour extraction method and device | |
CN103914699B (en) | A kind of method of the image enhaucament of the automatic lip gloss based on color space | |
CN106446779B (en) | Personal identification method and device | |
CN107578418A (en) | A kind of indoor scene profile testing method of confluent colours and depth information | |
CN103810750B (en) | Human body section ring based parametric deformation method | |
CN105069751B (en) | A kind of interpolation method of depth image missing data | |
CN102609684B (en) | Human body posture detection method and device | |
CN104408733B (en) | Object random walk-based visual saliency detection method and system for remote sensing image | |
CN108280397A (en) | Human body image hair detection method based on depth convolutional neural networks | |
CN104715238A (en) | Pedestrian detection method based on multi-feature fusion | |
CN106570476A (en) | Key size automatic extraction method in three dimensional human body measurement | |
CN106204223B (en) | image processing method and system | |
CN109636779A (en) | Identify the method, apparatus and storage medium of poultry volume ruler | |
CN102880870B (en) | The extracting method of face characteristic and system | |
CN107154058A (en) | A kind of method for guiding user to reduce magic square | |
CN105069816B (en) | A kind of method and system of inlet and outlet people flow rate statistical | |
CN105138975A (en) | Human body complexion area segmentation method based on deep belief network | |
CN107121140A (en) | A kind of location acquiring method based on Multiple Source Sensor | |
CN110189408A (en) | It is a kind of that the system and method for human body appearance data is obtained according to human body photo | |
CN106709122A (en) | Underwear designing system and underwear designing method | |
CN110032932A (en) | A kind of human posture recognition method based on video processing and decision tree given threshold | |
CN109446884A (en) | A kind of gait recognition method that the speed based on nonlinear kinetics index is unrelated | |
CN107818563A (en) | A kind of transmission line of electricity bundle spacing space measurement and localization method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20161228 |