| Automation technology is one of the important subjects of advanced garment manufacture technology in fashion design and development, production and marketing. The integrated program of digital fashion design and development is a whole process including3D anthropometric scanning, identification of body type and selection of data, intelligent fashion design, rapid pattern design, virtual fitting and so on. Independent studies on each topic have been constantly conducted by domestic and international study groups and individuals, but studies on the join of each part are new issues. For instance, the autofilter of body type data based on garment styles and the auto-import of rapid pattern design modules based on body type data. These are all key technique issues in the program of digital fashion design and development. For now before intelligent digital fashion design was achieved,2D working sketch is always used to express garment style. Therefore, extraction of technical parameters from working sketch and body type is a key issue of the transformation from garment sketch to pattern.Further, it is a technical bottleneck for the program of digital fashion design and development that extraction of technical parameters from garment illustration still totally depends on experience and manual operation. In order to join up each part of the program, and solve the problem of extraction technical parameters from garment illustration, make it independent of subjective judgment and achieve digital identification and extraction of parameters. Around this topic, an exploratory study was conducted in this paper.Firstly, a research was done to classify common garment illustrations, analyse and compare common fashion drawing with working sketch to learn their functions, features and applications. And an idea was put forward about setting up new kind of working sketch for computer recognition, as well as a solution including drawing rules, drawing tools and preparation of virtual model.Secondly, based on the working sketch for computer recognition, the image processing toolbox of MATLAB was used to identify the working sketch. By designing and coding, some typical technical parameters could be automatically identified from sketches, such as garment length and sleeve length in vertical as well as bust, waist and hip circumferences in horizontal.Thirdly, experiments were conducted to study the connection between garment circumferences and gaps between body and dress on the side for garment body type X and H. In the experiments,15garment styles were chosen for type X, and5were chosen for type H. Garments’ circumferences were varied from4cm to12cm with a range of5cm. By virtual fitting,175patterns were sewed virtually to set up sketches for identification. With the gap data, mathematical models were established to transform waist or hip gaps into circumferences for garment type X and H.Finally, an experement was conducted to test the function of the sketch identification program and the mathematical models. Three garments were chosen for the test, and by drawing directly according to the garment pictures, sketches were set up for identification. Patterns were made by the parameters extracted from sketches, and sewed virtually. The virtual fitting pictures were compared with the original garment pictures for similarity by subjective assessment. The scores proved the results of sketch identification were acceptable.In this study, identification of working sketch and extraction of technical parameters were explored which applied computer graphic for information extraction from working sketch. This brought up a new idea and solution for this area, and made a necessary step to solve the problem and get through the integrated program of digital fashion design. |
PDF Full Text Request