VISUAL CONTENT BASED IMAGE CLASSIFICATION USING FUNCTION, RULE AND TREE FAMILY DATA MINING CLASSIFIERS WITH LBG VECTOR QUANTIZATION METHOD

摘要

Importance of Content Based Image Classification is increasing in the field of Image Processing. Images are classified appropriately in their particular category using its features. This paper proposes LBG vector quantization approach for Image Classification with Function, Rule and Tree families of data mining classifiers. Here Experimentation is done using 7 codebook sizes and 11 data mining classifier. Total 77 variations are performed on a Test database. Data set of 500 images is used to compute the classification accuracy. 84.64% performance is observed by RBF Network classifier of Function Family with 1024 Codebook size of LBG.