The Evaluation of Fabric Prickle Based on BP Neural Network

Abstract:

A three-layer BP neural network model was established by relating subjective evaluation of fabric prickle level and 16 objective parameters from KES-FB system. The elastic gradient decrease method was adopted for network training to achieve the preset precision of the model which was later applied to fabric prickle level evaluation. Results from this method gave a considerably accuracy compared with actual subjective results which implied a compatibility between BP neural network and traditional subjective evaluation.

• Periodical: Advanced Materials Research (Volume 441)
• Pages: 645-650
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.441.645
• Online since: January 2012
• Authors: Tian Tian Li, Jian Zhong Shao, Jin Li Zhou, Tian Zuo Zhang
• Keywords: BP Neural Network (BPNN), Fabric Prickle, KES-FB System, Objective Evaluation

References

[1] L. Yi, Clothing comfort and its application, Textile Asia, 7 (1998) 29-33.

[2] J. Veitch, P. Kenins, G. R. S. Naylor, Proc 8th International Wool Res. Conf., 5 (1990) 286-295.

[3] J. W. Ou, Fabric softness evaluation with BP neural network, J. Shandong Textile Science & Technology, 24 (2009) 39-42.

[4] J. Guo, Neural network theory and realized by MATLAB 7, Electronics Industry Press, China (2005).

[5] J. D. Cao, The prediction of cotton fabric handle based on BP neural network, J. Textile Research, 6 (2003) 35-37.

[6] Research Center of Hunan Ramie Technology, The evaluation method of prickle sensation caused by wearing ramie garment, J. Linen Textile Technology, 20 (1997) 31-34.

[7] C. H. Chang, MATLAB neural network and application, National Defense Industry Press, China, (2005).