SUPERVISIED SEGMENTATION TECHNIQUE BASED ON MULTI-CHANNELS GABOR
S..M.Al..l and ISSRA J. MUIHSEN Remote Sensing Unit, College or Science, University or Baghdad Iraq, Baghdad, Al-Jaderiya
A bank or Gabor filters are used for segmenting special type of textured images. Directional filters are introduced to improve the output segmentation results. Texture features measure is proposed which based on subjecting the filtered images to a non-linear transformation and the “energy” within each image point. Different values nonlinear transform parameters are tried and used to perform the final segmentation Different smoothing windows are suggested and implemented to enhance the texture clustering in the nonlinear filtered images. A minimum distance classification algorithm is applied classify the featured images.
Keyword: Gabor filters, Image Segmentation, Image Filtering , Image Analysis.