Title :
Effect of gray-level re-quantization on co-occurrence based texture analysis
Author :
Patel, Mehul B. ; Rodriguez, Jeffrey J. ; Gmitro, Arthur F.
Author_Institution :
Dept. of ECE, Univ. of Arizona, Tucson, AZ
Abstract :
Gray-level co-occurrence matrices (GLCM) are widely used for texture analysis. The number of gray-levels used while computing GLCM is an important parameter but is often ignored. It is believed that the higher the number of gray-levels used, the better is the performance of the GLCM-based features. Contrary to this belief, we observed that using more gray levels than the actual range of pixel values in the image can give erroneous results. In this paper, we show how this occurs and discuss a way around this problem.
Keywords :
feature extraction; image texture; matrix algebra; GLCM-based feature; gray-level co-occurrence matrix; gray-level requantization; image pixel value; texture analysis; Algorithm design and analysis; Biomedical computing; Biomedical imaging; Computational efficiency; Fluorescence; Image analysis; Image texture analysis; Lighting; Pixel; Quantization; Gray-level co-occurrence matrix; histogram; re-quantization; texture analysis;
Conference_Titel :
Image Processing, 2008. ICIP 2008. 15th IEEE International Conference on
Conference_Location :
San Diego, CA
Print_ISBN :
978-1-4244-1765-0
Electronic_ISBN :
1522-4880
DOI :
10.1109/ICIP.2008.4711822
