Abstract: This invention relates to encoding and decoding images. The invention is a variant of the cascaded VQ with certain improvements acquired from the DCT and DWT approaches. The fundamental aspects of the invention are that codebooks are pre-processed when training them for predetermining the frequency distribution of the resulting codevectors, and each block is independently coded and decoded using a variable number of stages of difference coding needed for coding the particular block.

Abstract: The invention proposes a method of determining an index of a code vector (VCk) belonging to a predetermined lattice, characterised in that it includes the steps of:

Abstract: The present invention relates to a method of viewing and processing hyper-spectral image data compressed using a VQ algorithm. According to the invention the data is compressed using a codebook of codevectors including binary spectral vectors, which allows processing of the compressed data and viewing of data within a datacube without expanding the compressed data into the complete datacube. In order to view an image derived from a datacube, for each pixel within the image a location within the datacube is selected, an index value from the index map at that location is retrieved, and a spectral value from a spectral vector within the codebook is retrieved for said pixel, the spectral vector identified by the retrieved index. A single spectral vector is easily viewed by displaying a spectral vector from the codebook.

Abstract: The invention improves the refresh rate and minimizes the memory for display when reproducing a color image. A color image encoding device has a creation device which counts the number of colors used in color image data and creates a global palette which lists corresponding indexes when the number of colors is a specified value or less, a block division device which divides the color image data into a plurality of blocks, a local palette creation device which creates a local palette which lists indexes in the global palette when the number of colors in a block is less than the number of colors in the global palette, and a color index assigning device which assigns the indexes in the global palette to the input pixels when the number of colors is the same as the number of colors in the global palette, and assigns the indexes in the local palette to the input pixels when the numbers of colors is different from the number of colors in the global palette.

Abstract: A hyperspectral image encoder (10) for compressing hyperspectral imagery includes a differential pulse code modulation (DPCM) loop (26) to perform data decorrelation in the spectral domain and a discrete wavelet transform (DWT) processing means (28) to perform decorrelation in the spatial domain. The DPCM loop (26) determines an error image between a present image at the input of the encoder (10) and a predicted image. The DWT processing means (28) then divides the error image into a plurality of frequency subbands and quantizes information within the subbands in accordance with a plurality of predetermined quantization states to provide a coded output signal. In one embodiment, an interband predictor (24) is provided to predict an image, for use in calculating the error image, using the coded output signal.

Abstract: An image processing system including an image encoder and image decoding system is provided. The image encoder system includes an image decomposer, a block encoder, and an encoded image composer. The image decomposer decomposes the image into blocks. The block encoder which includes a selection module, a codeword generation module and a construction module, processes the blocks. Specifically, the selection module computes a set of parameters from image data values of a set of image elements in the image block. The codeword generation module generates codewords which the construction module uses to derive a set of quantized image data values. The construction module then maps each of the image element's original image data values to an index to one of the derived image data values. The image decoding system reverses this process to reorder decompressed image blocks in an output data file.

Abstract: A transformation unit 102 transforms an entered image to multi-resolution space. A quantizer 105 performs vector quantization on a local pattern of an image of a multi-resolution representation. A perspective-order calculating unit 107 extracts a plurality of code words, positions corresponding thereto and/or angle of rotation and/or scale, and the perspective-order relationship of a plurality of these representative vectors, from the quantized image. An algebraic encoder 108 encodes the input image based upon the extracted information. As a result, there are provided an image processing apparatus and method for asymmetric encoding without motion in three dimensions and extraction of a three-dimensional structure. Further, a transformation unit 1103 transforms an entered image to vector field, and a singularity detector 1104 detects a singularity in the transformed image.

Abstract: A linear transform apparatus for implementing a linear transform on input data values to produce linear transformed output data, the apparatus comprising: input means for inputting input data values one after another to each of a series of multiplication means; a series of multiplication means interconnected with the input means for multiplying a current input data value by a constant to produce a current multiplier output; an interconnection network interconnecting the series of multiplication means to predetermined ones of a series of signed accumulator means; a series of signed accumulator means each interconnected to the interconnection network, each of the signed accumulator means producing an intermediate accumulator output by accumulating a corresponding one of the current multiplier outputs with a corresponding previous intermediate accumulator output, each of the signed accumulator means outputting the intermediate accumulator output as a corresponding linear transformed output data value.

Abstract: Different images, or sub-images, are rendered at different wavelet decoding rates, the more rapidly decoded wavelets forming a focal region about which less detailed images, or sub-images, are formed. In a preferred embodiment, sets of images or sub-images form levels of a hierarchy, and the wavelet encodings of these images and sub-images are decoded at rates associated with each level of the hierarchy. A single image may be partitioned into sub-images, or regions, that form each level of the hierarchy, or independent images may be arranged in a hierarchical manner, and subsequently rendered using a hierarchy of wavelet decoding schemes. The hierarchy can be developed at a remote site, and a hierarchical wavelet encoding can be transmitted to a local site, so that the hierarchical foveation effect can be rendered while the details of the lower levels of the hierarchy are being communicated.

Abstract: A system is presented for compression of hyperspectral imagery. Specifically, DPCM is used for spectral decorrelation, while an adaptive 2-D discrete cosine transform (DCT) (22) coding scheme is used for spatial decorrelation. Trellis coded quantization (23) is used to encode the transform coefficients. Side information and rate allocation strategies (36) are discussed.

Abstract: A signal analysis system/method, for identifying the closest vector in a vector collection to a given input signal vector, comprising an input, a memory, and a processing unit. The memory stores a collection of vectors, and a table of mutual distances between pairs of the vectors in the collection. The processing unit may receive an input vector corresponding to the input signal.

Abstract: A method and apparatus for compressing a digitized image. Boundaries of regions within a digitized image are identified. Shapes are selected from a bank of shapes based on the identified boundaries to represent the regions. Position and size information based on the position and size of the regions within the digitized image are associated with the selected shapes. Values are transmitted indicating the selected shapes and the associated position and size information as a representation of the digitized image.

Abstract: Data is transmitted using multiple description vector quantization by first quantizing the source vector at a lattice vector quantizer. After quantization, a labeling function is applied to the quantized source vector, creating a plurality of data streams. Each data stream is encoded and transmitted over a separate channel. Furthermore, the encoded data is decoded by first retrieving a representation of a sublattice point from a set of sublattice points from the data stream and then, determining a single data code word being associated with the retrieved representation of the sublattice point.

Abstract: A method for data compression. An encoder receives data vectors from an original data set. The encoder uses a vector quantization codebook to encode the data vectors into encoded vectors. The codebook is constructed from a compound data set, where the compound data set includes real data vectors and artificial data vectors. The encoded vectors are indexed in the codebook and the indexes are transmitted across communication channels or transmitted to storage.

Abstract: A method of analysing a video signal for the presence of outlines to determine whether a frame of the video signal is likely to represent a cartoon is based on the concept that cartoons are often characterised by uniform colours within an image. In an improved version the signal is separated into signals representing smaller areas of the image before analysing each signal for the presence of uniform colours.

Abstract: Method, software, and apparatus (100, 300, 400) for performing rate controlled image compression are provided. A quantization matrix scaling factor that results in compression to a predetermined byte size is determined by first summing the absolute value of the difference of neighboring pixels selected from a plurality of rectangular regions of the image. The sums are then input to a model function that is evaluated to obtain the quantization matrix scaling factor. According to preferred embodiments the model function involves a dot product operation between a vector that includes the sums and a second vector that includes region weights.

Abstract: An information stream organized as a sequence of blocks is encoded according to a graceful degradation principle. The stream poses temporal non-uniform data processing requirements. In particular, with respect to a received block one or more blockwise defined control parameters and an associated blockwise processing load are detected. Under control of a processing load for one or more previous blocks, one or more later blocks before processing thereof get an adjustment of one or more of the control parameters. This lowers an expected load for the later block in case of an excessive load, and vice versa when detecting a sub-standard load.

Abstract: A picture target amount-of-code calculating unit calculates a target amount of codes for each picture type that corresponds to an encoding rate that has been determined by an audio/video encoding rate determining unit. An amount-of-code controlling unit controls the quantizing scale of a quantizing unit so as to make the target amount of codes for each picture type match with a total amount of codes of a unit as a whole. A unit address calculating unit calculates the addresses used when a search is made, as navigation data, on the basis of the target amount of codes for each picture type and the encoding rate. A unifying unit makes a construction of the unit while, on the other hand, a navigation data producing unit depicts navigation data at a foremost position of the unit.

Abstract: A method of decompressing image data. A VQ encoded image is received and decoded. Output image color space processing is performed in conjunction with the decoding. Output image color space processing may include color transformation, half-toning or both.

Abstract: A 2D image supplied from an image input unit including a wide view lens is sampled into a discrete form by an array sensor, and then mapped to a multi-resolution space by a 2D filter. The feature of the supplied image is detected, and then the mapped image is transformed to a local pattern about the detected feature, and then the coordinates of the position of the feature and the code word of the local pattern are formed into a set which is then encoded. The code is supplied to each cell of a stochastic automaton. The quantity of visual information is calculated in accordance with the quantity of mutual information between different cells of the stochastic automaton consisting of cells in blocks, the coordinates of the position of the feature and the distance from the feature to the optical axis so as to control the optical axis of the image input unit in such a manner that the quantity of visual information is maximized.

Abstract: A QIM watermark embedder uses a space filling quantizer to imperceptibly embed an auxiliary message signal into a media signal, such as an image, audio or video signal. A specific example of the space filling quantizer is a hexagonal lattice quantizer. A QIM watermark decoder uses a compatible space filling quantizer to decode the auxiliary message signal from a potentially distorted version of the watermarked media signal. The shape filling quantizer enhances the information carrying capacity of the digital watermark.

Abstract: A data compression system using vector quantization utilises a codebook or index tables constructed by finding a small set of points in an n-dimensional space which are representative of a much larger population. The space is partitioned such that the integrals of the probability density functions over all the subspaces are approximately equal and thereafter the centre of gravity of each subspace is taken to represent the subspace population. The partitioning is effected sequentially to map n-dimensional space into one dimensional code space.

Abstract: A picture encoding apparatus and the method in which the quantization step width of the lower hierarchy data having a resolution higher than that of the upper hierarchy data is determined for each predetermined block of respective hierarchy data based on the quantization step width determined by the upper hierarchy data having a low resolution, so that the additional code indicating the characteristics of quantizer can be omitted thereby the compression efficiency can be improved and the deterioration of picture quality can be reduced when picture data is hierarchical-encoded.

Abstract: Method of updating dictionaries in a data transmission system wherein strings of characters have to be transmitted in a compressed form from the transmit device to the receive device, the transmit device having a transmit dictionary storing codewords associated with the strings of characters which are transmitted instead of the strings of characters, the receive device having a receive dictionary storing codewords associated with the strings of characters, both dictionaries being updated each time a new string of characters has to be transmitted so that their contents remain identical.

Abstract: A system for classifying image elements comprising means for converting an image into a series of vectors and a hierarchical lookup table that classifies the vectors. The lookup table implements a pre-computed discrete cosine transform (DCT) to enhance classification accuracy. The hierarchical lookup table includes four stages: three of which constitute a preliminary section; the fourth stage constitutes the final section. Each stage has a respective stage table. The method for designing each stage table comprises a codebook design procedure and a table fill-in procedure. Codebook design for the preliminary stages strives to minimize a classification-sensitive proximity measure; codebook design for the final stage attempts to minimize Bayes risk of misclassification.

Abstract: Inverse quantization (36, 44, 60) of a two-dimensional array of video coefficients QF[v][u] is provided. A first weighting matrix {overscore (W)}[0][v][u]=2*W[0][v][u] for intra-coded macroblocks and a second weighting matrix {overscore (W)}[1][v][u]=2*W[1][v][u] for non-intra-coded macroblocks are pre-computed and stored (88). A quantization difference (quanti+1)to be used in selecting a weight factor is computed (70) based on a current macroblock quantization level. The weight factor is output from one of the weighting matrices in response to the quantization difference and a macroblock type identifier (66) (i.e., intra-coded or non-intra-coded). Subsequent weights Wi+1[w][v][u] are then computed (86) from a current weight Wi[w][v][u] and the weight factor.

Abstract: An encoding and/or decoding method and apparatus incorporates decoding commands into a compressed file. The decoding commands are instructions carried along to a decompressor and implement processing performed in the compressed domain. Then, the decompressor performs some processing steps to the data while decompressing it. This reduces the computation time for the imaging operations.

Abstract: A method and apparatus for optical character recognition particularly suitable for cursive and scripted text in one or more of several languages follows the tracings of the script and encodes them as a sequence of directional vectors. It reads a preprocessed word or sub-word of interconnected characters as a unit and the characters are accepted in a specific language only if all characters in a unit have been recognized by testing against a first set of language-specific rules without leaving a remainder of any vectors in the unit. If there are vectors remaining unused, it moves a character marker to utilize more or fewer vectors for the current sub-word in order to obtain recognition. If the vectors do not form a character in a first language, the invention consults a second set of language-specific rules and follows similar steps to obtain recognition of one or more characters against the second set of language-specific rules.

Abstract: A transmission method for video image data using an embedded bit stream in a hierarchical table-lookup vector quantizer comprises the steps encoding an image using hierarchical vector quantization and an embedding process to obtain an embedded bit stream for lossless transmission. The bit stream is selectively truncated and decoded to obtain a reconstructed image.

Abstract: A quantified symbol stream is sampled, using the quantified symbols prior to the current symbol, to generate a modeling value that characterizes the stream. The modeling value is used to generate, or retrieve, a probability function that provides the probability that the current symbol will be the next symbol in the symbol stream. The probability given by the probability function is then used to select the used to represent the data. The encoded symbol is later decoded using the characteristics of the code, and an identical probability function, to determine the value of the original quantified symbol.

Abstract: In a vector quantization of a digitized image, image regions (QB) to be quantized are selected smaller than, for example, image blocks (BB) of an image (B). In the reconstruction, image points (NBP) that are not located in an image region (QB) to be quantized are interpolated and/or extrapolated.

Abstract: An image compression system includes a vectorizer and a hierarchical vector quantization table that outputs embedded code. The vectorizer converts an image into image vectors representing respective blocks of image pixels. The table provides computation-free transformation and compression of the image vectors. Table design can be divided into codebook design and fill-in procedures for each stage. Codebook design for the preliminary stages uses a splitting generalized Lloyd algorithm (LBG/GLA) using a perceptually weighted distortion measure. Codebook design for the final stage uses a greedily-grown and then entropy-pruned tree-structure variation of GLA with an entropy-constrained distortion measure. Table fill-in for all stages uses an unweighted proximity measure for assigning inputs to codebook vectors. Transformations and compression are fast because they are computation free. The hierarchical, multi-stage, character of the table allow it to operate with low memory requirements.

Abstract: An image compression system includes a vectorizer and a hierarchical vector quantization table that outputs embedded code. The vectorizer converts an image into image vectors representing respective blocks of image pixels. The table provides computation-free transformation and compression of the image vectors. Table design can be divided into codebook design and fill-in procedures for each stage. Codebook design for the preliminary stages uses a splitting generalized Lloyd algorithm (LBG/GLA) using a perceptually weighted distortion measure. Codebook design for the final stage uses a greedily-grown and then entropy-pruned tree-structure variation of GLA with an entropy-constrained distortion measure. Table fill-in for all stages uses an unweighted proximity measure for assigning inputs a codebook vectors. Transformations and compression are fast because they are computation free. The hierarchical, multi-stage, character of the table allow it to operate with low memory requirements.

Abstract: A transmission method for video image data using an embedded bit stream in a hierarchical table-lookup vector quantizer comprises the steps encoding an image using hierarchical vector quantization and an embedding process to obtain an embedded bit stream for lossless transmission. The bit stream is selectively truncated and decoded to obtain a reconstructed image.

Abstract: A method and apparatus for compressing video data images uses optical processing techniques. The method and apparatus perform holographic optical correlation and apply holographic optical correlation in a feedback loop. A codebook of images or primitives for the correlation are stored in a holographic library.

Abstract: The invention relates to an image encoding/decoding method, apparatus thereof and a recording medium in which a program therefor is recorded, whereby the encoding/decoding can be obtained with high image quality at high speed. In an image encoding method which comprises producing a DC image composed of each block mean value by dividing an image data per B pixel into a block, making a part of said DC image a DC nest, and where the differential vector <dj> which is obtained by separating the DC value DCT from the pixel block <Rj> to be encoded is over an allowable value Z, calculating one or more orthogonal basis (&agr;k<Vk>), to which the differential vector is approximated, by the adaptive orthogonal transform (AOT) using the DC nest, each of the lowest n (n=log2 B) bits of base extraction blocks <Ui> which are down-sampled from the DC nest is set to 0.

Abstract: The memory requirement of MPEG decoders and SQTV/IQTV systems may be reduced by recompressing the MPEG decoded data stream before storing pixels in an external RAM. An efficient compression method for recompressing video picture data based on the tree-search vector quantization (TSVQ) is made more effective by optimizing the way the quantizer is chosen for quantizing the differences among adjacent pel vectors. This method is based on premultiplying a read-only table using quantized complexity measures relative to the centroids of the tree-like scheme used in the TSVQ processing. A plurality of precalculated tables of quantization of the prediction error of a physical parameter of blocks of digital data are produced. For each one of the regions in which a block is divided, the calculated and quantized complexity measure provides an address that selects the most appropriate precalculated table for quantizing the prediction error.

Abstract: A method and apparatus for compressing digital image data to improve the efficiency of serial data transmission is disclosed. More specifically, the present invention accomplishes image transmission by providing multiple image processing modules, and then selecting the module that will process the entire in the most efficient manner, based upon the content of the data contained in the image. Image content is analyzed using hierarchical vector quantization.

Abstract: The invention relates to a device for the verification of time-dependent, user-specific signals which includes means for generating a set of feature vectors which serve to provide an approximative description of an input signal and are associated with selectable sampling intervals of the signal; means for preparing an HMM model for the signal; means for determining a probability value which describes the probability of occurrence of the set of feature vectors, given the HMM model, and a threshold decider for comparing the probability value with a threshold value and for deciding on the verification of the signal. For better adaptation of the verification to the relevant user and the user-specific signals assigned thereto, it is proposed to provide means for the automatic person-specific evaluation of the relevance of the features collected in the feature vectors by way of an LDA transformation and for the automatic selection of features evaluated as relevant.

Abstract: A method and apparatus for compressing and decompressing texture patterns using a predefined two or more stage codebook with stored texture block vectors. The method of compressing includes selecting a texture block in the uncompressed texture pattern and determining a distortion value of the selected texture block as compared to a texture block vector stored in the two or more stage codebook. The distortion determination is repeated for a different texture block stored in the two or more stage codebook, until the determined distortion value is the smallest for all texture block vectors stored in the two or more stage codebook. The indices of the stored texture block in the two or more stage codebook are retrieved and stored. The retrieved indices are stored in an index block within an index map, wherein the index blocks correspond to texture blocks in the texture pattern. The two or more stage codebook is generated from a representative texture pattern.

Abstract: In an image processing method and apparatus, image data having multi-value levels for one pixel is input, and the input image data is quantized such that an output area of one pixel is adapted to an output device in which an output area of one pixel changes depending on the position of the pixel. A quantizing process executes an arithmetic operation based on an algorithm of a neural network on the basis of a value obtained by multiplying an output value by a weight corresponding to an area of each pixel. Therefore, even if pixels have different maximum luminances, the different numbers of bits, and different color expression capabilities, an optimum half-tone process can be performed by an algorithm based on a cellular neural network, and a high-quality image can be obtained.

Abstract: A pattern matching coding device for conducting coding of a unit pattern based on pattern matching with a library pattern, which device, as to a mark pattern determined to be matching, when the number of match error pixels of the pattern is not more than a prescribed value, codes, without coding the bit map, an identification ID of a matching bit map in a library and an identification flag indicating that the number of match error pixels is not more than the prescribed value.

Abstract: A method and apparatus for efficiently transmitting digital image data is disclosed. More specifically, the present invention accomplishes fast data transmission by analyzing the content of input data, and then retrieving data that closely matches that which would be produced if the data were subjected to some form of data compression from storage. The retrieved data is transmitted to the retrieving device, thereby eliminating the need for very time consuming data compression processes.

Abstract: A Two Row Buffer Image Compression (TROBIC) technique is provided that operates on 32×2 blocks of pixels and uses a block-based coding scheme to provide alternative coding modes for image blocks containing different types of image data. The TROBIC algorithm automatically evaluates the image presented to determine regions containing text, graphics, synthetic images and natural images along with determining the most effective combination of global-index coding, local-index coding, lossless coding and wavelet coding to use to guarantee the desired compression ratio and high quality. The wavelet coding technique, in addition to being part of TROBIC, may be used independently in a wide variety of compression applications. The wavelet coding technique is carried out by a simple block codec with wavelets (SBCW) that uses wavelet filters to transform 32×2 image blocks into subband coefficients organized in a tree structure.

Abstract: Pixel blocks of an input image are type classified based on an analysis of pixel values for each respective pixel block. A discrete cosine transform (DCT) is performed on the pixel values of each pixel block, and a quantization modification process thresholds and/or quantizes the resulting DCT coefficients based on the type classification of the respective pixel block. Once the coefficients are modified in this way and encoded, the resulting data can be decoded and dequantized in compliance with the standard JPEG sequential mode data syntax in order to construct a perceptually faithful representation of the image, without passing any additional information to the decoder concerning the quantization modification.

Abstract: A system, method, and software product provide for unsupervised identification of complex, nonlinear subspaces in high dimensional data. The system includes a vector quantization module, a weighted topology representing graph module, and an encoding module. The vector quantization module takes vector data inputs and extracts a group of inputs about a number of cluster centers, using a globally optimized clustering process. The weighted topology representing graph module creates a weighted graph of the vector space, using the cluster centers as nodes, weighting edges between nodes as a function of the density of the vectors between the linked nodes. The encoding module uses the weighted graph to recode the input vectors based on their proximity to the cluster centers and the connectedness of the graph. The recoded vectors are reinput into the vector quantization module, and the process repeated until termination, for example at a limited number of cluster centers.

Abstract: An image compression system includes a vectorizer and a hierarchical vector quantization table that outputs embedded code. The vectorizer converts an image into image vectors representing respective blocks of image pixels. The table provides computation-free transformation and compression of the image vectors. Table design can be divided into codebook design and fill-in procedures for each stage. Codebook design for the preliminary stages uses a splitting generalized Lloyd algorithm (LBG/GLA) using a perceptually weighted distortion measure. Codebook design for the final stage uses a greedily-grown and then entropy-pruned tree-structure variation of GLA with an entropy-constrained distortion measure. Table fill-in for all stages uses an unweighted proximity measure for assigning inputs to codebook vectors. Transformations and compression are fast because they are computation free. The hierarchical, multi-stage, character of the table allow it to operate with low memory requirements.

Abstract: An image compression system includes a vectorizer and a hierarchical vector quantization table that outputs embedded code. The vectorizer converts an image into image vectors representing respective blocks of image pixels. The table provides computation-free transformation and compression of the image vectors. Table design can be divided into codebook design and fill-in procedures for each stage. Codebook design for the preliminary stages uses a splitting generalized Lloyd algorithm (LBG/GLA) using a perceptually weighted distortion measure. Codebook design for the final stage uses a greedily-grown and then entropy-pruned tree-structure variation of GLA with an entropy-constrained distortion measure. Table fill-in for all stages uses an unweighted proximity measure for assigning inputs to codebook vectors. Transformations and compression are fast because they are computation free. The hierarchical, multi-stage, character of the table allow it to operate with low memory requirements.

Abstract: A compression process of dividing each word to be compressed into least and most significant bits, and applying a different compression ratio to each. If the compression processes are lossy, both parts of the output can also be losslessly compressed before being output. Hierarchical vector quantization can be issued for either lossy compressor.

Abstract: An image compression system includes a vectorizer and a hierarchical vector quantization table that outputs embedded code. The vectorizer converts an image into image vectors representing respective blocks of image pixels. The table provides computation-free transformation and compression of the image vectors. Table design can be divided into codebook design and fill-in procedures for each stage. Codebook design for the preliminary stages uses a splitting generalized Lloyd algorithm (LBG/GLA) using a perceptually weighted distortion measure. Codebook design for the final stage uses a greedily-grown and then entropy-pruned tree-structure variation of GLA with an entropy-constrained distortion measure. Table fill-in for all stages uses an unweighted proximity measure for assigning inputs to codebook vectors. Transformations and compression are fast because they are computation free. The hierarchical, multi-stage, character of the table allow it to operate with low memory requirements.