An efficient single-pass adaptive bandwidth compression technique using the discrete cosine transform is described. The coding process involves a simple thresholding and normalization operation on the transform coefficients. Adaptivity is achieved by using a rate buffer for channel rate equalization. The buffer status and input rate are monitored to generate a feedback normalization factor. Excellent results are demonstrated for coding of color images at 0.4 bits/pixel corresponding to real-time color television transmission over a 1.5 Mbit/s channel.

This paper presents the development and evaluation of a visual texture feature extraction method based on a stochastic field model of texture. Results of recent visual texture discrimination experiments are reviewed in order to establish necessary and sufficient conditions for texture features that are in agreement with human discrimination. A texture feature extraction technique involving autocorrelation function measurement of a texture field, combined with histogram representation of a statistically decorrelated version of the texture field, is introduced. The texture feature extraction method is evaluated in terms of a Bhattacharyya distance measure.

Earlier experiments by Julesz et al. on visual discrimination are reviewed and extended in the context of texture analysis of natural images. Stochastic methods are developed for the generation of pairs of synthetic texture fields possessing multiple gray levels, significant spatial correlation, and joint moments or probability densities of controllable form. Results of several visual discrimination experiments with these computer generated texture fields are then presented. It is demonstrated, for the stochastic models investigated, that humans cannot effortlessly discriminate between pairs of spatially correlated texture fields with differing third-order probability densities when their lower order densities are pairwise equal. This is a further verification of Julesz's conjecture. Also, it is shown that human observers are sensitive to relatively small changes in the spatial autocorrelation function. Examples are presented of discriminable texture fields with identical means, variances, autocorrelation functions, and third-order nearest neighbor moments.

New computational techniques for pseudoinverse and Wiener image restoration are presented. These techniques are based upon the modification of an observed image array so that the numerical model for the image blur contains a circulant matrix blur operator. Utilization of Fourier transform processing techniques then permits scalar computation, as contrasted to vector computation required for standard processing. Several image restoration examples are given.

Two-dimensional transform coding and hybrid transform/DPCM coding techniques have been investigated extensively for image coding. This paper presents a theoretical and experimental extension of these techniques to the coding of sequences of correlated image frames. Two coding methods are analyzed: three-dimensional cosine transform coding, and two-dimensional cosine transform coding within an image frame combined with DPCM coding between frames. Theoretical performance estimates are developed for the coding of Markovian image sources. Simulation results are presented for transmission over error-free and binary symmetric channels.

A new unitary transform called the slant transform, specifically designed for image coding, has been developed. The transformation possesses a discrete sawtoothlike basis vector which efficiently represents linear brightness variations along an image line. A fast computational algorithm has been found for the transformation. The slant transformation has been utilized in several transform image-coding systems for monochrome and color images. Computer simulation results indicate that good quality coding can be accomplished with about 1 to 2 bits/pixel for monochrome images and 2 to 3 bits/pixel for color images.

The classical signal processing technique known as Wiener filtering has been extended to the processing of one-and two-dimensional discrete data by digital operations with emphasis on reduction of the computational requirements. In the generalized Wiener filtering process a unitary transformation, such as the discrete Fourier, Hadamard, or Karhunen-Loéve transform is performed on the data that is assumed to be composed of additive signal and noise components. The transformed data is then modified by a filter function, and the inverse transformation is performed to obtain the discrete system output. The filter function is chosen to provide the best mean square estimate of the signal portion of the input data.

This paper develops performance criteria for evaluating transform data coding schemes under computational constraints. Computational constraints that conform with the proposed basis-restricted model give rise to suboptimal coding efficiency characterized by a rate-distortion relationR(D)similar in form to the theoretical rate-distortion function. Numerical examples of this performance measure are presented for Fourier, Walsh, Haar, and Karhunen-Loève transforms.

This paper considers the effects of channel errors on the digital transmission of color images. In the transmission system analyzed, each picture element of a color image is represented by three color components. Two cases are considered: the tristimulus representation and the luminance and chromaticity representation. Each color component is coded with a constant-word-length PCM code and passed through a binary symmetric channel. Channel errors are related to chromaticity and luminance errors in a reconstructed image by the use of color space diagrams. Statistical characteristics of image errors are determined for various color representations. It has been found that if the tristimulus values of a color are transmitted over the channel, the effects of channel errors are to desaturate ali colors toward a mid-gray. For luminance/chromaticity representations, channel errors tend to shift the hue of transmitted colors toward cyan, magenta, or yellow.

a matrix factorization is presented for a Hadamard matrix of order twelve that permits a Hadamard transform of this order to be computed with substantially fewer operations than by simple matrix multiplication. The matrix factorization is extended to Hadamard matrices of order 2nX 12 where n is an integer.

This bibliography contains books, papers, and articles on the subject of television (TV) bandwidth reduction. Also included is related information on the processing and coding of video signals, and the subjective evaluation of TV images.