SOME INVESTIGATIONS ON SIGNAL PROCESSING TRANSFORMS

Abstract

In recent years there has, been a. growing intrest regarding the application of orthogonal transforms to digital signal and image processing. This primarily owing to the impact of high speed digital computers and the rapid advances in digital technology and consequent development of special purposes digital processors. Orthogonal transforms have been used. to process various types of data from speech, radar, biological, and biomedical sources. and data from satellite television pictures. A data can be efficiently transmitted over noisy, narrow bandwidth channels by transmitting a proper orthogonal transform of the data. An inverse transform at the receiver resconstructs the data.Suitable transformation must leave the entropy of the data source unchanged but place the data in a form less sensitive to channel errors and more amenable to bandwidth reduction. Since most of the compression results from dropping the transform coefficients with smaller energy, it is.desirable to have a transform which compacts most of the energy in as few Transform components as possible. Transforms that have •proven usefullness include the discrete Fourier, discrete cosine, Haar, Hadamard, Slant and Karhunen Loeve......