Protection of images that are encoded using subband coding from channel error is addressed. In this scheme the low-pass subband is encoded using DPCM (differential pulse-code modulation), and the other subbands are encoded using a scalar quantizer. The quantizers are all Lloyd-Max quantizers, from which the representation levels have fixed length codewords. First, considering only single errors in each codeword, a channel error distortion measure is derived for each quantizer, that is, for each subband. Codewords are assigned to the quantizer representation levels, yielding a low value of the distortion measure. Next, sets S/sub ij/ consisting of the jth bit from subband i are formed. Each set S/sub ij/ is assigned a particular BCH code C/sub ij/. An algorithm that optimally assigns BCH codes C/sub ij/ to each set S/sub ij/, based on a channel error distortion measure for the entire image, is derived. The protection scheme is adaptive, because each set of bits within each subband can be assigned a different error protection code. Examples show that this approach is preferable to assigning equal error protection codes to each set of bits. It is shown that in the case of a channel error probability of 10/sup -3/, only 5% to 10% extra bits are needed for adequate channel error protection.>

Necessary and sufficient conditions are given for block codes to be capable of correcting up to t/sub 1/ symmetric errors, up to t/sub 2/ unidirectional errors, and up to t/sub 3/ asymmetric errors, as well as detecting from t/sub 1/+1 up to d/sub 1/ symmetric errors that are not of the unidirectional type, from t/sub 2/+1 up to d/sub 2/ unidirectional errors that are not of the asymmetric type, and from t/sub 3/+1 up to d/sub 3/ asymmetric errors. Many known conditions on block codes concerning error correction and/or detection appear as special cases of this general result. Further, some codes turn out to have stronger error correcting/detection capabilities than they were originally designed for.>

An algorithm is presented for the generation of binary DeBruijn sequences using feedback shift registers (FSRs). The algorithm is based on the principle of cycle joining, through which all the cycles in the cycle structure of a FSR can be joined together to one complete cycle, thereby producing a DeBruijn sequence of period 2/sup n/, where n is the length of the FSR. By a proper choice of the feedback function of the FSR O(2/sup 2n///sup log2n/) DeBruijn sequences of period 2/sup n/ can be generated, requiring only 3n bits of storage and at most 4n FSR shifts for the generation of the next bit in the sequence.>

A hierarchical blur identification method which is able to identify severe blurs having a relatively large support size is proposed. The method is based on the expectation-maximization algorithm and begins by identifying a low-pass version of the point-spread function (PSF) from a reduced-resolution version of the blurred image. The result is used to initialize the identification of the PSF from a higher-resolution image. By repeating this process, the support size and resolution of the identified PSF are gradually increased, achieving an efficient identification of large PSFs.>

An algorithm is proposed that estimates displacement vectors and segments each frame into three regions: unchanged, predictable, and unpredictable. The segmentation result is used for the adaptation of the causal window and the initial estimate, both essential estimation information. The algorithm consists of a two-stage detector where the stages are separated by a Wiener-based displacement estimator. The detectors are efficient, suboptimal, maximum a posteriori estimators of special first-order 2-D Markov random-field models. The detectors are shown to have a substantially improved performance compared with detectors previously used, and the prediction error is reduced compared with existing schemes. Initial experiments indicate good coding efficiency.>

The authors present a maximum-likelihood blur identification method, which estimates the required parameters from the observed noisy blurred image itself, and use the expectation-maximization algorithm to solve the resulting complicated problem of optimizing the likelihood function. A priori information about the unknown parameters in the form of initial conditions and parametric image and blur models are incorporated to make the algorithm applicable to realistic blurs and to improve the identification results. Experimental results are presented on a 256 pixel*256 pixel synthetically blurred by a 2-D Gaussian point spread function with various standard deviations. The approach results in a flexible iterative algorithm that is computationally far more efficient than directly optimizing the likelihood function.>

The authors investigate the application of subband coding (SBC) to the progressive transmission of images. The main objective of progressive transmission is to allow the receiver to recognize a picture as quickly as possible at minimum cost. After this it can be decided to receive further picture details or to abort the transmission. SBC turns out to be very suitable for such a system, because the low-pass version of the image is easily recognized for any image type. Further details of the picture are obtained by sequentially receiving the (encoded) higher bandpass versions of the image. Because SBC has good data compression properties, not many bits are needed to transmit the low-pass image and the other subbands. Comparisons to other progressive transmission schemes, such as 8*8 block DCT (discrete cosine transform) encoding, show that SBC is favorable, both in mean-squared error and in a subjective sense.>

A brief introduction is given on the theory of codes correcting unidirectional errors, in the context of symmetric and asymmetric error-correcting codes. Upper bounds on the size of a code of length n correcting t or fewer unidirectional errors are then derived. Methods in which codes correcting up to t unidirectional errors are constructed by expurgating t-fold asymmetric error-correcting codes or by expurgating and puncturing t-fold symmetric error-correcting codes are also presented. Finally, tables summarizing some results on the size of optimal unidirectional error-correcting codes which follow from these bounds and constructions are given.>

Source coding and channel coding are embedded in a video coding algorithm for low bitrates (64 kb/s) to obtain a complete encoder which is resilient against channel errors. The bits retained for the channel coder are spent for synchronization words, error correction, and update of erroneous parts in the image. The performance of the complete codec is evaluated by simulations in which channels with different characteristics are modeled.>

An optimal bit allocation algorithm is presented that is suitable for all practical situations. Each source to be coded is assumed to have its own set of admissible quantizers (which can be either scalar or vector quantizers) which do not need to have integer bit rates. The distortion versus rate characteristic of each quantizer set may have an arbitrary shape. The algorithm is very simple in structure and can be applied to any practical coding scheme (such as a subband coder) that needs dynamic bit allocation.>

A novel two-dimensional subband coding technique is presented that can be applied to images as well as speech. A frequency-band decomposition of the image is carried out by means of 2D separable quadrature mirror filters, which split the image spectrum into 16 equal-rate subbands. These 16 parallel subband signals are regarded as a 16-dimensional vector source and coded as such using vector quantization. In the asymptotic case of high bit rates, a theoretical analysis yields that a lower bound to the gain is attainable by choosing this approach over scalar quantization of each subband with an optimal bit allocation. It is shown that vector quantization in this scheme has several advantages over coding the subbands separately. Experimental results are given, and it is shown the scheme has a performance that is comparable to that of more complex coding techniques.>

Upper bounds to the maximum number of codewords in a binary code of length n and asymmetric distance Delta are derived for some values of n and Delta . A method is given in which a code of length n-m and asymmetric distance at least t+1 is constructed by expurgating and puncturing a code of length n and Hamming distance at least 2t+1. Novel asymmetric error-correcting codes are constructed by applying this method to some celebrated symmetric error-correcting codes. a table is presented on the size of optimal asymmetric error-correcting codes of length less than 24 and asymmetric distance less than 6.>

This paper presents a form of sub-band coding of images in which the advantages of both DPCM and vector quantization are combined in a single subband coder. A 16 band decomposition of the image frequency band is employed using a tree structure of two-dimensional separable quadrature mirror filters and decimators. The 16 parallel sub-bands created in this way are regarded as a 16-dimensional vector process and coded as such using a vector equivalent of DPCM. Here each vector is predicted as a single quantity and its residual is coded using vector quantization, this vector quantizer being nested within the vector-DPCM loop. As will be shown this technique yields a good SNR performance in the bit rate region between 0.4 and 0.7 bits per pixel, both for images inside and outside a large training set.

New upper bounds on the size of codes correcting asymmetric errors are derived by sharpening some of the constraints in the integer programming problem of Delsarte and Piret. It is shown that their code for length9and asymmetric distance2is optimal.

In this paper we extend the transform domain oriented estimation algorithm introduced in [1] in which the calculation of the displacement vector was obtained from the transform domain coefficients. The performance of the algorithm is verified within a hybrid coding configuration. In this paper only transform domain block matching algorithms are considered. The block-match procedure makes use of the displacement matrix H defined in [1]. A matrix decomposition method is described in order to show that a practical implementation is very well possible. The properties of the translation invariant matrices are explained by using the ordered Walsh Hadamard transform as an example. The procedure however enables the use any other orthogonal transform. An important issue with respect to the hardware complexity of this motion compensated hybrid coder is the use of only one transform. The performance of the proposed new algorithm is shown and a video tape containing a very critical videoconferencing scene (i.e. split screen and a hard switch to full screen with heavy motion) will be presented. The sequences are coded at a bitrate of 384 kbit/s and 64 kbit/s respectively. Results of the compensation in the pixel domain and in the transform domain are compared and it is shown that the latter approach results in an improved performance of the estimation procedure.