Varieties Classification into Plain, Patterned and Un-patterned from Fabric Images

Abstract

The presented work gives a methodology to classify fabric images as plain, patterned and un-patterned. Discrete Wavelet Transform is applied and wavelet features are extracted. Feed Backward Selection Technique is used in the feature selection phase. Two prediction models, namely, Support Vector Machine and Artificial Neural Network are used. The overall classification rates of 81% and 86.5% are obtained for fabric types using Support Vector Machine and Artificial Neural Network classifiers. The classification rates for varieties of non-plain fabric images are found to be 84% and 88% respectively.