Image Classification using Fuzzy Block Truncation Coding

This paper gives a new method for image feature extraction named "Fuzzy Block Truncation Coding". Performance of Fuzzy Block Truncation Coding for classification with different methods of pre-processing, in various color spaces is analyzed. Fuzzy Block Truncation Coding is used to extract features from images in different color spaces such as RGB, YUV, YCrCb, HSV and LUV. Different classifiers such as C4.5, Naive Bayes and Random Forest with different pre-processing methods such as Fuzzy and PKIDiscretize are used. Experimental results obtain using standard dataset proves that proposed feature extraction scheme works better image classification.