Authors: M.D. Shieh, C.C. Cheng, Y.N. Sun
Addresses: Department of Industrial Design, National Cheng Kung University, 1 University Road, Tainan, 701 Taiwan, ROC. ' Amtek SYSTEM CO., LTD., Taipei, ROC. ' Department of Computer Science and Information Engineering, National Cheng Kung University, 1 University Road, Tainan, 701 Taiwan, ROC
Abstract: A fabric retrieval system, trained by the Back-Propagation Network (BPN), has been implemented. The new implementation of the retrieval system focuses primarily on the similarity in judgement of a human being|s Kansei for fabric images. The colour composition information is quantised and used to index the colour features of each fabric image. The colour features and the corresponding Kansei features are trained using a Neural Network (NN) model. The new similarity modelling approach, which integrates colour features and a human being|s perception features into a NN training model, can be further extended to costume fashion design, product styling design, etc.
Keywords: fabric retrieval; neural networks; ANNs; Kansei; colour features; similarity modelling; human perception; costume fashion; fashion design; product styling; product design; fabric design; textile design; content-based retrieval; fabric images.
International Journal of Simulation and Process Modelling, 2007 Vol.3 No.4, pp.219 - 228
Available online: 16 Dec 2007 *Full-text access for editors Access for subscribers Purchase this article Comment on this article