This paper introduces a statistical image model based on occlusion and maximum entropy. The statistical model combines a fundamental property of image formation, occlusion, with both object-image shape and nonuniform object-image intensity. The model is a composition of individual object-images that have random positions, shapes, and intensities, and that occlude both background and one another. We derive the autocorrelation and second-order probability density functions of this model and give several examples.

We present an algorithm that combines image segmentation and motion field estimation. The segmentation includes the occluded and uncovered background regions, the self-occluded and uncovered object regions, and the common moving regions of the objects.

This paper introduces a new realization for maximum likelihood parameter and time-delay estimation using two channels. We obtain the result by using coupled Karhunen-Loeve expansions to derive the likelihood function. The new realization illuminates the analytical relationship between maximum likelihood time-delay estimation and other methods. Experimental results that support the theoretical findings are included.

This paper describes novel adaptation strategy for hybrid transform/DPCM image coders. This strategy has yielded a 2-8 dB increase in signal-to-noise ratio in the transmission of a head and shoulders image at bit rates of 1-4 bits per pel and one-dimensional transform block sizes of 2-16.

This paper generalizes time-discrete autoregressive source coding results of rate-distortion theory to two dimensions. A 2-D discrete autoregressive source is defined and shown to produce a 2-D widesense Markovian field. The rate distortion function of the source is then obtained under assumption of Gaussian field statistics and a squared error fidelity criterion. A procedure for generating an ensemble of 2-D codewords whose statistics satisfy the variational equations forR(D)is given. These 2-D codewords are, by space-time mappings, 1-D tree codes, and it is noted that a tree coding theorem of Jelinek, Berger, Davis and Hellman applies. The problem of instrumenting nearly optimum 2-D sequential encoding is discussed briefly. The paper stresses potential application to image coder design.

This paper shows that, under certain weak restrictions, a two-dimensional discrete Markov process can be represented either "causally" by a one-sided difference equation, or "noncausally" by a multiple-sided difference equation. The former representation is pertinent to the analysis of image coders and processors that operate sequentially on scanned image data. The latter representation is applicable to the analysis of block coders and processors. The general relation between the two representations is given and the specialized forms resulting under separable scene covariance studied.

This paper defines anM-ary generalized likelihood ratio test (MGLRT) that overcomes Root's early objection to the application of generalized likelihood ratio testing to the resolution of correlated signals. The proposed test extends the form of a conventional binary generalized likelihood ratio test (GLRT) in a manner that permits a generalization of the minimax properties of the binary test to theM- hypotheses case. When the estimated signals are orthogonal, the test reduces to a sequence of conventional binary tests duplicating the performance of a narrow-band matched filter envelope-detector receiver.

With the availability of the computer for two-dimensional picture processing, digital restoration of deterministically degraded sampled images is a practical reality. This correspondence presents a mathematical formulation of discrete modulation transfer compensation of sampled imagery. The analysis makes no assumption regarding the band limitedness of the imaging system and accounts explicitly for the effects of spectrum foldover.

The problem of reconstructing a finite-duration finite-energy signal that has been band limited and sampled is considered. An interpolation formula is derived that, in principle, permits perfect signal reconstruction in the noiseless case provided only that the sampling frequency exceeds the cutoff frequency of the band-limiting filter. The degradation introduced by measurement noise on the samples is evaluated.