Research on Characteristic Properties of Gray Level Co-Occurrence Matrix

Abstract:

Gray level co-occurrence matrix (GLCM) is a second-order statistical measurement. In order to understand the characterization degree of GLCM’s different feature properties, we use images of Brodatz texture images as experimental samples, analyze the change process of feature properties in horizontal, vertical and principal and secondary diagonal directions under the situation of some elements’ dynamic changes such as distance of pixels pair, size of moving window and gray level quantization,. By analyzing the experimental results, this paper can provided certain referential significance in how to select feature properties reasonable in the application of image retrieval and classification and identification which are based on using GLCM as feature.

• Periodical: Applied Mechanics and Materials (Volumes 204-208)
• Pages: 4755-4759
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.204-208.4755
• Online since: October 2012
• Authors: Ying Chen, Feng Yu Yang
• Keywords: Change Curve, Feature Property, Grey Level Co-Occurrence Matrix, Texture

References

[1] H. Tamura, S. Mori, T. Yamawaki: IEEE T. Syst. Man Cy. B., Vol. 8 (1978), pp.460-473.

[2] R. Chellappa, S. Chatterjee: IEEE T. Acoustic, Vol. 33 (1985), pp.959-863.

[3] M. Comer, E. Delp: IEEE T. Image Process, Vol. 8 (1999), pp.408-20.

[4] F. Liu, R.W. Picard: IEEE T. Pattern Anal., Vol. 18 (1996), pp.722-733.

[5] L.M. Kaplan: IEEE T. Image Process, Vol. 8 (1999), pp.1572-1585.

[6] R.M. Harlick: P. IEEE, Vol. 67 (1979), pp.786-804.

[7] The USC-SIPI Image Database [2012-2-28]http: /sipi. usc. edu/database.

[8] L.H. Yuan, L. Fu, Y. Yang, J. Miao: J. Comp. Appl., Vol. 29(2009), pp.1018-1821.