Share Email Print

Proceedings Paper

A transform (MRT) naturally suited to directional pattern analysis
Author(s): Rajesh Cherian Roy; R. Gopikakumari
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Transforms having directional properties are important in image processing applications. M-dimensional Real Transform (MRT), having frequency and phase indices, is computed using real additions only. In this paper, the directional nature of MRT is presented. Each MRT coefficient is formed from combination of image coefficients in well-defined patterns, and this resembles a spatial filtering operation. The specific filter structure depends on frequency, and location of the filter on image lattice depends on phase. Combining the patterns associated with MRT coefficients having the same frequency index, global patterns can be created over the entire image lattice. These global patterns can be used as pattern strength indicators in images containing directional patterns. A subset of global patterns of a 16x16 MRT is used to estimate the orientation field of fingerprint images. Results obtained show that MRT can be used for orientation estimation.

Paper Details

Date Published: 28 January 2013
PDF: 6 pages
Proc. SPIE 8760, International Conference on Communication and Electronics System Design, 876023 (28 January 2013); doi: 10.1117/12.2012433
Show Author Affiliations
Rajesh Cherian Roy, Federal Institute of Science and Technology (India)
R. Gopikakumari, Cochin Univ. of Science & Technology (India)

Published in SPIE Proceedings Vol. 8760:
International Conference on Communication and Electronics System Design
Vijay Janyani; M. Salim; K. K. Sharma, Editor(s)

© SPIE. Terms of Use
Back to Top