DocumentCode :
2639047
Title :
3D Fashion Fast Modeling from Photographs
Author :
Li, Wei-Long ; Lu, Guo-Dong ; Geng, Yu-Lei ; Wang, Jin
Author_Institution :
State Key Lab. of CAD&CG, Zhejiang Univ., Hangzhou, China
Volume :
1
fYear :
2009
fDate :
March 31 2009-April 2 2009
Firstpage :
738
Lastpage :
742
Abstract :
The paper presents a technique of 3D fashion modeling from photographs of wearing clothes in front and back views. Firstly, an efficient segmentation method is applied on photographs to obtain the silhouette of the fashion. Then, a template-based feature extraction algorithm is introduced to determine the feature points on the garment. Finally, a view-dependent deformation technique is described to construct the fashion by deforming the garment template. Our segmentation algorithm is derived from mathematical morphology and image difference method. The deformation technique is related to free-form deformations and vector field of mannequin. With our deformation method, the main feature of fashion is preserved. Compared with other predefined fashion modeling approaches, the efficient and realistic of constructing is greatly increased. The functionality of garment model constructed by our method can apply to some others applications for garment industry.
Keywords :
clothing industry; feature extraction; image segmentation; image texture; mathematical morphology; photography; production engineering computing; 3D fashion fast modeling; garment industry; image difference method; mathematical morphology; photograph; segmentation method; template-based feature extraction algorithm; view-dependent deformation technique; Biological system modeling; Clothing; Computer science; Deformable models; Feature extraction; Flowcharts; Humans; Image segmentation; Morphology; Paper technology;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer Science and Information Engineering, 2009 WRI World Congress on
Conference_Location :
Los Angeles, CA
Print_ISBN :
978-0-7695-3507-4
Type :
conf
DOI :
10.1109/CSIE.2009.838
Filename :
5171272
Link To Document :
بازگشت