Research On Evaluation Model Of Clothes Fit Based On Individual Virtual Mannequin

In the context of rapid development of Internet of Things, evaluation model of clothes fit based on individual virtual mannequin has been studied in this paper, which includes data acquisition, transmission and storage of 3D body, acquiring and analysis parameters of individual virtual mannequin, modeling of individual virtual mannequin, modeling of virtual clothes, evaluation model of clothes fit, etc.The first work is application of data mining technology to analysis 3D scanning data of women’s torso that is extracted from 3D body data of point cloud, which is acquired by means of [TC]2 system.3D cloud point data that has been cleaned is the data source of data mining. The key points and key planes are obtained by application of descriptive algorithm of data mining. The other points can be obtained by application of forecasting algorithm of data mining. The error analysis has been done.Based on the conclusion of data mining, the data acquisition system of 3D body based on multi-photo has been developed. The body contours is extracted from images that are captured at designated angle. Image processing and distortion correction must be done in advance. The characteristic points and key points are extracted from body contours, based on which the other points are calculated.3D body data are acquired through the above steps. A feasible solution of representation and transmission of 3D body data is put forward by the application of XML technology according to association rules that have been extracted by data mining. XML documents are interpreted and verified in the environment of Visual Studio 2005. Database model of 3D body is built and implemented in MS SQL Sever 2005 database.The precondition of individual virtual mannequin is to obtain parameters that can embody the individual features of virtual mannequin. In order to meet the demand of individual virtual mannequin modeling in the network environment, a solution of getting parameters for modeling individual virtual mannequin is put forward by the application of Web Service technology. In the last place, model of individual virtual mannequin is built through tri patch modeling techniques. In addition, many lines of reference such as neck line, bust line, waist line, hip line, front center line, back center line, side seam are built on the model, which is prepared for evaluation model of clothes fit.In the end of this paper, evaluation model of clothes fit is built, which consists of evaluation model based on size designation, evaluation model based on reference lines and evaluation model based on surface of body according to complexity from low to high. The evaluation model of clothes fit has the applicable scope to be broad, which can meet the requirement of the development E-commerce, fashion online marketing and Internet of Things. The evaluation model is highly efficient and the three sub-models are also the progressive process in which clothes that don’t even match can be weeded out, thus the three sub-models are not completing execution which not only save the resources and calculation, but also shorten the time of evaluation. During the procedure of implementation, sub-model is optional for user, which is flexible and can meet all kinds of requests from different users. Information of evaluate result is converted into graphic by using visualization technology. The implement of visual communication between people and data or among people makes the evaluation model have user friendly interface and feasible.The following work has been innovatively completed in this paper.1.3D scanning data analysis of women’s torso based on data miningAfter 3D scanning data of women’s torso is analyzed by means of descriptive algorithm of data mining, two conclusions can be drawn. In storey dimension, women’s torso is divided into 50 layers evenly spaced from bottom to top. Layer 10, 20,29,31,34,36 are both the maximum value of first order difference and the minimum value of second order difference. Layer 10,20,31 are severally located near the hip line, waist line and bust line. Layer 29,31 are respectively situated near the under bust line and upper bust line. Layer 34 is seated between upper bust line and bust line. The six layers are the key layers that present surface features of women’s torso. In fan dimension, women’s torso is equally divided into 8 portions by three via origin planes that the frontal direction contour, the lateral direction contour and the rotation 45°direction contour respectively seat. The Angle degree of two nearby planes is 45°. The three group fans that drop shadow to plane of projection are the key fans that present surface features of women’s torso. After 3D scanning data of women’s torso is analyzed by means of forecasting algorithm of data mining, the following conclusion can be drawn. There is a multi valued dependency relationship between key points and other points on each layer. The value of a non-key point can be predicted according to the value of the three nearby key points by using linear regression data mining algorithm. The error is in the acceptable range.2. Building standards system of data acquisition, transmission and storage of 3D bodyThe method of 3D body data acquisition based on the conclusion of 3D body data mining is studied which is by means of three photographs that are captured at designated angle. The acquisition and distortion of body image has been adjusted in order that data accuracy can fulfill the employ of clothing industry. The body contours is extracted from images that are captured at designated angle and have been corrected the distortion. The characteristic point and key points are extracted from body contours, based on which the other points are calculated.3D body data are acquired through the above steps. Accessory appliances such as table turning of angular dimensioning have been fabricated. The data acquisition system of 3D body that can provide point cloud data has been developed. A feasible solution of representation and transmission of 3D body data is put forward by the application of XML technology according to association rules that have been extracted by data mining. XML documents are interpreted and verified in the environment of Visual Studio 2005. Database model of 3D body is built and implemented in MS SQL Sever 2005 database.3. Building individual virtual mannequin that is used for evaluation model of clothes fitIndividual virtual mannequin that is used for evaluation model of clothes fit is built. The parameters to build individual virtual mannequin are provided in four levels that are size level, characteristic point level, key point level and cloud point level in the order from simple to complex. The parameters of different levels can be combined. The third level named key point level is adopted in this paper. The scheme of acquiring parameter for individual virtual mannequin is built through the use of Web Service technology, which can realize interconnection of loosely coupling distributed component. Both various data structure stored on different operating systems and different hardware platforms of various merchants and diverse program interface of individual body data are availably integrated, which share the individual body data of clients in all merchants.3D coordinate data of point cloud of women’s torso are created according to key point data. Surfaces model of body based on Cubic B-spline technology is built after point cloud data being grid processed. In addition, many lines of reference such as neck line, bust line, waist line, hip line, front center line, back center line, side seam are built on the model, which is preparing for evaluation model of clothes fit.4. Building three levels evaluation model of clothes fitIn the architecture of three levels evaluation model of clothes fit, the first level named evaluation model based on size designation is the minimum level, in which algorithm is easy, the number of data required is small, operation time is less, which is appropriate for evaluation of clothes that are loose in style and have great flexibility in fabric. The second level is evaluation model based on reference lines, which is based on the first level. Only if evaluate result of the first level is provisional accreditation and user selects the option of next level, evaluation model based on reference lines is carried out. The third level is evaluation model based on surface of body, which is based on the second level. Only if evaluate result of the second level is provisional accreditation and user selects the option of next level, evaluation model based on surface of body is carried out. The third level is the maximum level, which needs many detailed data of both body and clothes. Operation time is more. The third level is appropriate for evaluation of clothes that are fitted in style and have small flexibility in fabric. The three levels evaluation model of clothes fit has the applicable scope to be broad, which can meet the requirements of the development E-commerce, fashion online marketing and Internet of Things. The evaluation model is highly efficient and the three sub-models are also the progressive process in which clothes that don’t even match can be weeded out, thus the three sub-models are not completing execution which not only save the resources and calculation, but also shorten the time of evaluation. During the procedure of implementation, sub-model is optional for user, which is flexible and can meet all kinds of requests from different users. Information of evaluate result is converted into graphic by using visualization technology. The implement of visual communication between people and data or among people makes the evaluation model have user friendly interface and feasible.