extended faculty

Mysore Raghuveer

Raghuveer Rao, Ph.D.
Associate Professor

Ph.D., Electrical Engineering
University of Connecticut

2123 Gleason
E-Mail: mrreee@rit.edu

Research Highlights

1) Wavelet Transform: Image Processing Applications - The wavelet (WT) transform is ideally suited for imaging applications that involve processing the image at various levels of resolution and/or processing texture information. A problem that we are working on (which has yielded a crop of several more interesting problems) matches a wavelet given object as closely as possible. Performing a WT with this matched wavelet provides excellent pattern recognition capability when the object size is allowed to vary in a scene (due to zooming, for example).

Another application being investigated is the decomposition of an image into two images where one retains the textural information and the other contains the luminance information. These images are further processed for enhancement application such as, for example, compensating intensity fall off.

Finally, application of the WT for processing color images is being investigated.

2) Digital Image Halftoning - Halftoning is the process by which a rendering of a continuous tone image is provided using a small number of display or printing levels. Image display is a key component of multimedia information systems and halftoning is likely to assume more importance in the context of multimedia efforts. Our research has yielded image halftoning procedures based on the concept of tracking taken from control systems. An interesting variation which has yielded excellent tone and edge reproduction is the noise thresholding approach. Research is continuing in color halftoning, compression of halftoned images, application of wavelet transforms to halftoning, etc.

3) Image Processing with Higher-Order Spectral Techniques - Higher order spectra, which are generalizations of the common power spectrum, contain information about the Fourier transform phase and magnitude. Furthermore, they possess information about nonlinear transformations and are less sensitive to various types of noise. Our research has exploited these properties in applications such as the reconstruction of moving objects in noise, restoring speckle degraded images and coded aperture imaging. We have also developed algorithms for phase and magnitude reconstruction, and detection and estimation of nonlinear interactions in random media.

Selected Publications

Raghuveer, M.R. ``Time Domain Approaches to Estimating Quadratic PhaseCoupling,'' IEEE Trans. on Automatic Control---mini-Special issue on Higher-Order Statistics in Syst Theory and Signal Processing, 48-56, January 1990.

Dianat, S.A. and M.R. Raghuveer, ``Fast Algorithms for Phase and Magnitude Reconstruction from Bispectra,'' Optical Engineering, May 1990.

Sundaramoorthy, G., M.R. Raghuveer and S.A. Dianat,``Bispectral Reconstruction of Signals in Noise: Amplitude Recovery Issues,'' IEEE Trans. on Acoust., Speech and Sig. Proc., 1297-1306, July 1990.

Song, Jin , S. Wear and M.R. Raghuveer, ``Restoration of Speckle-Degraded Images using Bispectra,'' J. Opt. Soc. Amer A, pp.371-376, March 1992.

Samar, V.J., K. Swartz and M.R. Raghuveer, "Multiresolution analysis of event related potentials,'' to appear in Brain and Cognition, April 1995.