The Spectral Characterizing Model Based on Optimized RBF Neural Network for Digital Textile Printing
Digital textile printing was appeared in 1980s, which printed directly onto the surface of fabric with digital original by ink-jet printer and got high quality color printing textile. The spectral characterization of printer is the key technology for the textile printing. This paper presented a spectral characterizing model based on RBF neural network, which optimized the RBF neural network by extending the input variables of neural network. Experimental results showed that the 90% spectral errors of testing color samples are less than 0.04, the average spectral error is 0.025, the maximum spectral error is 0.066; while the 90% color differences (ΔE2000) of testing color samples are less than 2.8, the average value is 1.89, the maximum value is 8.5. That means the model could effectively improve the characterization chromaticity and spectral precision.
KeywordSpectral characterization Spectral model Optimized neural network Digital textile printing
- 2.Zhou R, Yu Z (2010) Unscrambling the development status-quo of international nonwoven equipment. China Text Leader 9:70–72Google Scholar
- 3.Aehwal WB (2002) Textile chemical principles of digital textile printing (DTP). Colour ge 12:33–34Google Scholar
- 9.Wan XX, Liu Q (2014) Review of spectral printer characterization. J Image Graph 19(7):985–997Google Scholar
- 11.da Cruz LF, Freire RZ, Reynoso-Meza G et al (2016) RBF neural network combined with self-adaptive mode and genetic algorithm to identify velocity profile of swimmers. In: Proceedings of 2016 IEEE Symposium Series on Computational Intelligence (SSCI), pp 1 – 7Google Scholar