Abstract: A multiple description (MD) joint source-channel (JSC) encoder in accordance with the invention encodes n components of a signal for transmission over m channels of a communication medium. In illustrative embodiments, the invention provides optimal or near-optimal transforms for applications in which at least one of n and m is greater than two, and applications in which the failure probabilities of the m channels are non-independent and non-equivalent. The signal to be encoded may be a data signal, a speech signal, an audio signal, an image signal, a video signal or other type of signal, and each of the m channels may correspond to a packet or a group of packets to be transmitted over the medium. A given n×m transform implemented by the MD JSC encoder may be in the form of a cascade structure of several transforms each having dimension less than n×m. The transform may also be configured to provide a substantially equivalent rate for each of the m channels.

Abstract: A multiple description (MD) joint source-channel (JSC) encoder in accordance with the invention encodes n components of an image signal for transmission over m channels of a communication medium. In an illustrative embodiment which uses statistical redundancy between the different descriptions of the image signal, the encoder forms vectors from transform coefficients of the image signal separated both in frequency and in space. The vectors may be formed such that the spatial separation between the transform coefficients is maximized. A correlating transform is then applied, followed by entropy coding, grouping as a function of frequency, and application of a cascade transform. In an illustrative embodiment which uses deterministic redundancy between the different descriptions of the image signal, the encoder may apply a linear transform, followed by quantization, to generate the multiple descriptions of the image signal.

Abstract: Interpolating filter banks are constructed for use with signals which may be represented as a lattice of arbitrary dimension d. The filter banks include M channels, where M is greater than or equal to two. A given filter bank is built by first computing a set of shifts &tgr;i as D−1 t1, i=1, 2, . . . M−1, where ti is a set of coset representatives taken from a unit cell of the input signal lattice, and D is a dilation matrix having a determinant equal to M. A polynomial interpolation algorithm is then applied to determine weights for a set of M−1 predict filters Pi having the shifts &tgr;1. A corresponding set of update filters Ui are then selected as Ui=Pi*/M, where Pi* is the adjoint of the predict filter Pi. The resulting predict and update filters are arranged in a lifting structure such that each of the predict and update filters are associated with a pair of the M channels of the filter bank.

Abstract: A multiple description (MD) joint source-channel (JSC) encoder in accordance with the invention encodes n components of an audio signal for transmission over m channels of a communication medium, where n and m may take on any desired values. In an illustrative embodiment, the encoder combines a multiple description transform coder with elements of a perceptual audio coder (PAC). The encoder is configured to select one or more transform parameters for a multiple description transform, based on a characteristic of the audio signal to be encoded. For example, the transform parameters may be selected such that the resulting transformed coefficients have a variance distribution of a type expected by a subsequent entropy coding operation. The components of the audio signal may be quantized coefficients separated into a number of factor bands, and the transform parameter for a given factor band may be set to a value determined based on a transform parameter from at least one other factor band, e.g.

Abstract: Interpolating filter banks are constructed for use with signals which may be represented as a lattice of arbitrary dimension d. The filter banks include M channels, where M is greater than or equal to two. A given filter bank is built by first computing a set of shifts &tgr;i as D−1ti, i=1, 2, . . . M−1, where ti is a set of coset representatives taken from a unit cell of the input signal lattice, and D is a dilation matrix having a determinant equal to M. A polynomial interpolation algorithm is then applied to determine weights for a set of M−1 predict filters Pi having the shifts &tgr;i. A corresponding set of update filters Ui are then selected as Ui=Pi*|M, where Pi* is the adjoint of the predict filter Pi. The resulting predict and update filters are arranged in a lifting structure such that each of the predict and update filters are associated with a pair of the M channels of the filter bank.

Abstract: The invention provides improved processing of image data and other types of signal data by representing the signal data in such a way that, when separated into packets, all packets are of approximately the same importance. As a result, if some of the packets are, for example, randomly lost in a lossy packet network, the resulting degradation in a reconstructed version of the signal is substantially uniform regardless of which packets are lost. A given image or other signal may be separated into packets using an energy equalization process, a signal whitening process, or other suitable technique. The invention may be utilized in conjunction with data transmission over a wide variety of lossy packet networks, and may be used with raw or compressed image data, as well as video, audio, speech and other types of signal data. Moreover, the invention can be used to facilitate the processing of signal data in multiresolution storage applications.

Abstract: Interpolating filter banks are constructed for use with signals which may be represented as a lattice of arbitrary dimension d. The filter banks include M channels, where M is greater than or equal to two. A given filter bank is built by first computing a set of shifts .tau..sub.i as D.sup.-1 t.sub.i, i=1,2, . . . M-1, where t.sub.i is a set of coset representatives taken from a unit cell of the input signal lattice, and D is a dilation matrix having a determinant equal to M. A polynomial interpolation algorithm is then applied to determine weights for a set of M-1 predict filters P.sub.i having the shifts .tau..sub.i. A corresponding set of update filters U.sub.i are then selected as U.sub.i =P.sub.i */M, where P.sub.i * is the adjoint of the predict filter P.sub.i. The resulting predict and update filters are arranged in a lifting structure such that each of the predict and update filters are associated with a pair of the M channels of the filter bank.

Abstract: Non-uniform filters are designed with a method which utilizes an orthogonal subbanding technique. A tiling representation of a signal is established. The signal is localized for both time and frequency using a mixed tiling representation of the signal. A prototype is then established using the joint time frequency representation of the signal. Lastly, non-uniform filters are developed utilizing the prototype.

Abstract: A method and apparatus for encoding and decoding multi-dimensional signals is disclosed. Non-separable multi-dimensional filters are utilized. Analysis filters encode the signal by applying a windowing function and a reduction function to the multi-dimensional signal. This produces a prototype. The prototype is then replicated as many times as required by utilizing a modulation function. Synthesis filters then decode the encoded signal by utilizing the inverse of the encoding process.

Abstract: Method and apparatus are disclosed for deinterlacing of an interlaced video frame sequence using interpolation estimations, such as spatial and temporal interpolations. Interpolations requiring a less accurate estimation of missing pixel values in the frames being deinterlaced, such that an interpolation may be performed with a minimum of error, are performed before interpolations which require a more accurate estimation of missing pixel values for performing an interpolation, such that estimates of missing pixel values are obtained with a minimum of error. Interpolation estimations are weighted in combination for computing approximations of missing pixel values in accordance with the errors associated with the respective interpolations.

Abstract: A method and apparatus for reducing correlated errors in subband coding systems with quantizers is disclosed. A subband coding system comprises a plurality of subband analysis filters to divide the frequency spectrum of the input signal into subbands, individual subband quantizers for coding each subband by a preselected number of quantization levels, corresponding subband decoders and subband synthesis filters. The transfer function of each of the subband synthesis filters is advantageously determined based on the transfer functions of the subband analysis filters as well as on the characteristics of the quantizer used to code the corresponding subband. Specifically, the synthesis filter transfer functions may be based on a perfect reconstruction filter bank or a quadrature mirror filter bank, as well as on the gain factors of a gain plus additive noise linear model for the Lloyd-Max quantizers used to code the corresponding subbands.