A NOVEL APPROACH COMBINING GABOR WAVELET AND MOMENTS FOR TEXTURE SEGMENTATION
K. Muneeswaran, Lambodaran Ganesan, S. S. Arumugam, P. HARINARAYAN · International Journal of Wavelets Multiresolution and Information Processing · 2005
In this work, an effective method has been proposed for texture segmentation, which incorporates the best features of filter bank and statistical approaches. This technique combines the features of Gabor wavelets (filter based) and General Moments (statistical) approaches. The method has been successfully tested for various textures from Brodatz texture collection. The relative performance of this method against the conventional approaches has been analyzed using Fisher Criterion.