Accession Number : ADA189570

Title :   Parameterization of Mottle Textures.

Descriptive Note : Technical rept.,

Corporate Author : ROCHESTER UNIV NY DEPT OF COMPUTER SCIENCE

Personal Author(s) : Brown, C M ; Hinkelman, Elizabeth ; Jain, Sanjay

PDF Url : ADA189570

Report Date : Jun 1987

Pagination or Media Count : 33

Abstract : Parameterization of textures can be useful for detection of textual similarities and matching. In this project we have developed stochastic model to generate a set of parameters from the texture image domain and frequency domain. This model is aimed at quantification of textures for detection of similarities differences. Our attention has been concentrated on the parameterization of mottled textures. To test the model we have used it to generate texture images back from the parameters obtained from image analysis. The similarities and differences between the generated image and original images are used to refine and test the parametric model. Keywords: Textures, Parameterization, Stochastic model, Image domain, Texture synthesis, Texture classification, Power spectrum.

Descriptors :   *CLASSIFICATION, *IMAGE PROCESSING, *TEXTURE, *PARAMETRIC ANALYSIS, DETECTION, FREQUENCY, IMAGES, MATHEMATICAL MODELS, PARAMETRIC ANALYSIS, POWER SPECTRA, STOCHASTIC PROCESSES, SYNTHESIS

Subject Categories : Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE