Abstract: Shape information is scalably encoded and decoded by interleaved scan line (ISL) and raster scan line (RSL) methods and the encoded and decoded shape information is used for encoding texture information. The shape information of a chrominance (UV) component is encoded to compensate for the chrominance (UV) component. The encoding method is independently applied to the shape and texture component of each block. A scalable encoder of a still image using wavelets compresses pixels by using the characteristics between ISL pixels in a shape of the layer to be encoded or pixels between two layers in encoding the shape between layers. It is therefore possible to restore the shape and texture at high speed by performing the scalable encoding according to resolution, such as for searching for an image in a database/library. Also, tile coding a large image field to independently restore desired parts is fast and efficient.

Abstract: A method comprising receiving media data in a compressed, digital domain, analyzing motion vectors associated with the received media content while in the compressed digital domain and identifying one or more objects in the received media data based, at least in part, on the motion vector analysis.

Abstract: A method and apparatus for efficiently encoding images using a set of non-orthogonal basis functions, thereby allowing reduction of file size, shorter transmission time, and improved accuracy. The non-orthogonal basis functions include homogenous color basis functions, luminance-encoding basis functions that have luminance edges and chromatic basis functions that exhibit color opponency. Some of the basis functions are non-orthogonal with respect to each other. Using these basis functions, a source vector is calculated to provide a number of coefficients, each coefficient associated with one basis function. The source vector is compressed by selecting a subset of the calculated coefficients, thereby providing an encoded vector. Because the method is highly efficient, the image data is substantially represented by a small number of coefficients. In some embodiments, the non-orthogonal basis functions include two or more classes. A wavelet approach can also be utilized.

Abstract: A compressed bitstream is scaled down to a reduced rate bitstream by first demultiplexing a compressed input bitstream to extract video objects as elementary input bitstreams having a first bit rate. A transcoder converts each elementary input bitstream to an elementary output bitstream having a second bit rate. The first bit rate is less than the second bit rate. A transcoding control unit, coupled to the transcoder, supplies control information for the transcoder. A multiplexer composes the elementary output bitstreams into a compressed output bitstream having the second bit rate.

Abstract: In an apparatus for transcoding a compressed video, a generator simulates constraints of a network and constraints of a user device. A classifier is coupled to receive an input compressed video and the constraints. The classifier generates content information from features of the input compressed video. A manager produces a plurality of conversions modes dependent the constraints and content information, and a transcoder produces output compressed videos, one for each of the plurality conversion modes.

Abstract: A decoder decodes encoding binary image data on each block in order to recover a target block of the binary image. The image decoder includes motion compensated blocking which obtains a reference block from a reference binary image by applying motion compensation using motion information. A selector selects a statistical model from among a plurality of statistical models, based on states of pixels surrounding a reference pixel in the reference block, in which the reference pixel also corresponds to a target pixel in the target block. An arithmetic decoder recovers the target block by decoding the encoded data using the selected statistical model. In one embodiment, the pixels surrounding the reference pixel in the reference block are pixels positioned within one pixel distance from the reference pixel.

Abstract: Methods and apparatuses for still image compression, video compression and automatic target recognition are disclosed. The method of still image compression uses isomorphic singular manifold projection whereby surfaces of objects having singular manifold representations are represented by best match canonical polynomials to arrive at a model representation. The model representation is compared with the original representation to arrive at a difference. If the difference exceeds a predetermined threshold, the difference data are saved and compressed using standard lossy compression. The coefficients from the best match polynomial together with the difference data, if any, are then compressed using lossless compression. The method of motion estimation for enhanced video compression sends I frames on an “as-needed” basis, based on comparing the error between segments of a current frame and a predicted frame.

Abstract: An object of this invention is to provide a coding method and a coding apparatus for efficiently coding both a pixel value signal and a shape signal which are included in an image signal to be coded, wherein a prediction changing unit outputs a reference pixel value changing signal and a reference shape signal to a switching circuit which selects a reference pixel value signal and to a switching circuit which selects a reference shape signal, respectively, so as to control the switching circuit in a way that appropriate reference signals are to be selected.

Abstract: An efficient motion estimation technique for an arbitrarily-shaped video object reduces the number of searches for motion estimation for shape coding and texture coding. The technique can be easily employed in an MPEG-4 encoder for coding Video Object Planes (VOPs). In a pixel loop, an m×m block traverses a reference video image with the video object, and the alpha plane value of each pixel is examined. If the alpha plane values in a block are not all the same, this indicates the block overlaps the object boundary. Accordingly, a shape coding mask value for the coordinate of a reference pixel in the block, such as the top, left pixel, is set to “1”. As the block traverses the reference video image, a shape coding mask that generally follows the shape of the object boundary is developed to define a shaped reference image search area. A texture coding mask is defined similarly when the blocks overlap or are inside the object.

Abstract: A subjective evaluator delineates regions of an information space to be encoded in a qualitatively preferential or non-preferential manner such that the encoded information space comprises one or more of normal, emphasized or de-emphasized information content.

Abstract: The present invention provides a transmission apparatus for video information that allows wireless transmission of video information at a low rate regardless of differences in OSs.

Abstract: A coding protocol provides for coding video data that has been organized as video objects. The protocol provides a keyregion to permit coding of a region of data within the video object having common attributes. According to the protocol a keyregion is identified by a keyregion header, which includes a resync marker that uniquely identifies the keyregion header, a keyregion position signal indicating an origin and a size of the keyregion; and data of the common attribute. Data following the keyregion is coded according to the common attribute.

Abstract: A bi-level image, coded as transition data indicating positions of transitions in individual dot lines, is analyzed by comparing transition positions occurring in a consecutive group of dot lines, and detecting segments in different dot lines in the group that have the same level and join up in the direction parallel to the dot lines. The line density of the bi-level image can be reduced by replacing the group of dot lines with a single new dot line that includes the transition positions of all dot lines in the group, except for interior transition positions of joining segments. The line density can be increased by copying transition data of each dot line to a new dot line, adjusting the transition positions of segments that join segments in an adjacent dot line.

Abstract: A method for wavelet encoding image-representative signals, including the following steps: identifying shape information of a shape in the image; within rows of pixels of a shape, performing wavelet decompositions for segments of consecutive pixels in the shape; for segments of pixels having out of the shape pixels in a wavelet decomposition, forming a zero tree that includes don't-care nodes and having uncoded don't-care descendents of don't-care nodes.

Abstract: There is provided an image decoding apparatus for decoding a coded image signal having an image identifier based on a data structure of the image signal. This apparatus comprises a data analyzer for analyzing the coded image signal with reference to the image identifier to decide whether the coded image signal is a coded arbitrary shape signal including both of coded shape bit streams obtained by coding a shape signal which represents the shape of each object as one of the components of a display image and coded pixel value bit streams obtained by coding a pixel value signal representing the gradation of the object, or a coded binary signal including only coded shape bit streams obtained by coding a shape signal representing a display image of binary information; a first decoder for decoding the coded shape bit streams by a first decoding process; and a second decoder for decoding the coded pixel value bit streams by a second decoding process.

Abstract: There is disclosed a macroblock-based segmentation and background mosaicking method for object-based coding of image sequence having a stationary background. An image comprising a background having rare change and an object having slow change taken from a position and focus fixed camera is segmented into changing regions and stationary regions in units of macroblock of 16×16 pixels that is most widely used as a basic coding unit in a video coding method or apparatus. A background mosaic is constructed with only macroblocks corresponding to the regions having rare temporal change. The background mosaic is newly constructed with respect to an entire image whenever there are any changes in the image sequence, thereby maximizing coding efficiency. Therefore, the present inventions realizes a very low complexity, thus facilitating real time operation and maximizing efficiency in is object shape coding.

Abstract: An image recognition method comprising an extraction step of extracting local feature elements in an input image; a conversion step of arranging the local feature elements extracted at said extraction step on each of cells of predetermined lattice pattern composed of a plurality of cells having predetermined sizes and figures while substantially maintaining relative position relationship between a plurality of extracted local feature elements which are near each other; and a matching step of estimating a resemblance between a plurality of input images by matching the plurality of local feature elements arranged on the lattice pattern at said conversion step.

Abstract: A system, method and computer program product are provided for compression of characters. Upon receipt of a plurality of characters, at least one test is performed on the characters to determine if an outline of the characters is capable of being approximated by a mathematical approximation. If the outline of the characters is capable of being approximated, the characters are approximated using the mathematical approximation.

Abstract: A generic spatially scalable shape encoding apparatus and method for deriving shape information for chrominance components from luminance component. The present generic spatially-scalable shape encoding applies a series of subband (e.g., wavelet) filters to obtain N-levels of wavelet decomposition for the texture information of both luminance and chrominance components. The application of the corresponding subsampling filters of said subband filters is applied in a manner such that the shape of the chrominance can be derived from the shape of the luminance at the same spatial layer.

Abstract: An apparatus and method for obtaining scalability of a video object (VO) whose position and/or size changes with time. The position of an upper layer picture and that of a lower layer picture in an absolute coordinate system are determined so that corresponding pixels in an enlarged picture and in the upper layer picture may be arranged at the same positions in the absolute coordinate system.

Abstract: A method, system, and machine-readable medium for classifying an image element as one of a plurality of categories, including assigning the image element based on a ratio between an unoccluded perimeter of the image element and an occluded perimeter of the image element and coding the image element according to a coding scheme associated with the category to which the image element is classified. Exemplary applications include image compression, where categories include image foreground and background layers.

Abstract: Moving picture transforming system which executes transformation of moving picture coding scheme at high speed and with less degradation of picture quality, in that there are provided a motion compensating means 42 for compensating an image based on the motion information output from the variable length decoding means; a moving picture decoding means 120 consisting of an image memory for storing a reference image used by the motion compensating means; a motion information memory portion 20 for storing the motion information output from the moving picture decoding means; a motion prediction controlling portion 60 for outputting a candidate of the motion vector and controlling execution of the motion prediction, based on the motion information accumulated in the motion information memory portion; a motion candidate memory 53 for storing a plurality of outputs of the motion prediction controlling portion in sequence; an adaptive motion predicting means 51 for performing the motion prediction of the image output

Abstract: An image encoding method includes a step in which image data of components other than image data of components representing the transparency of an image is encoded from among image data of components which includes image data of components representing the transparency of the image, and a step in which an image file is created from image data of components encoded in the encoding step and image data of components other than image data of components encoded in the encoding step. Accordingly, because image data of components representing the transparency of the image is not encoded in the encoded image file, the image shape can be recognized. Moreover, because the image data of the remaining components is encoded, the security of the details of the image data can be maintained.

Abstract: In a method and a system for encoding and transmission of still images having at least one region of interest (ROI), the ROI coefficients of an image transformed into the frequency domain, preferably using a wavelet transform are encoded so that they are transmitted first and can be decoded by a receiver without transmission of the boundary of the ROI. In a preferred embodiment the coefficients belonging to the ROI are shifted so that the minimum ROI coefficient is larger than the largest background coefficient. A receiver can then perform an opposite procedure and thereby obtain the ROI. By specifying how much the coefficients need to be shifted in order to avoid sending shape information several advantages are achieved. Thus, it is possible to avoid sending shape information and to avoid shape encoding at encoder side. Furthermore, there is no need for a shape decoder at receiver side, and there is no need for the receiver to produce the ROI mask.

Abstract: The present invention is intended to exactly extract a feature-region along its contour line without erroneously extracting any moving object other than the feature-region. An extraction signal from a feature-region main-portion extracting circuit is sent to a small region eliminating circuit whereby an erroneously extracted small region is detected and eliminated as an erroneously extracted noisy portion. The small region eliminating circuit compares a small region with four detection patterns, discriminates it as a region other than the feature (face)-region if no match is found and causes a center pixel of a window of 3×3 pixels to have a logical value of zero. An output of the small region eliminating circuit is sent to a blank filling circuit which compares its input with four detection patterns, discriminates it as a face-region if a match is found and makes a central pixel of a window of 13×13 pixels have a logical value of 1.

Abstract: It is an object to encode an image signal or a shape signal more efficiently than the prior art. As a means to accomplish the object, the change pixel detector 2 receives the input signal 1 as an input signal and detects the pixel which changes the two-valued pixel value. Further, the change pixel predictor 4 also reads out the reference image stored in the memory 3, and predicts the change pixel of the particular input signal. The difference value calculator subtracts the output of the change pixel predictor 4 from the output of the change pixel detector 2. The difference value rounder 7 compares the tolerance value e and the prediction error D, and outputs x which requires the minimum bit number for being encoded in the value D−e≦x≦D+e. The output of the difference value rounder 7 is encoded by the decoder 8 to become the encoded signal 9.

Abstract: Disclosed is a scheme for creating, manipulating and transmitting images of a 3-D object through networks. A useful layering coupled with vector quantization allows an efficient combination of compressing and encoding of both the topology (connectivity) and geometry (vertex coordinates) of polygon meshes such as triangles. Connectivity information is contained in both generalized triangle strips and exception strips, the latter characterized by triangles for which all three vertices lie in the parent vertex layer. Geometry encoding is achieved with a productive vector quantization scheme which uses correction vectors for which distortion is merged into a following correction vector. Progressive connectivity transmission is achieved by using both intra-layer and inter-layer decomposition and reconstruction. The process encodes connectivity and geometry separately. Encoding of 3-D objects having left-right mirror symmetries using the process is described.

Abstract: An image coding method is provided, in which the method includes the steps of: extracting edge information which represents an edge part of an original image; obtaining density information of an edge smoothed image from the original image by smoothing the edge part; obtaining coded edge information by coding the edge information according to a first coding algorithm; obtaining coded density information by coding the density information of the edge smoothed image according to a second coding algorithm; and sending the coded edge information and the coded density information as coded information to an image decoding apparatus.

Abstract: A linear transformation of the image signal on a segment with an arbitrary geometrical shape includes the following steps: the sub-division of the segment into sub-segments of regular shapes (rectangular, square or linear shapes), the application of an orthogonal linear transformation to each sub-segment, the combining of the coefficients coming from the first transformation into classes of coefficients according to a predetermined criterion, and the iteration of the transformation on the classes of coefficients, and which can be applied to the processing of images for encoding.

Abstract: The invention disclosed here is a method for an encoder to encode audiovisual information for transmission to the decoder without any prior knowledge of the computational capabilities of the decoder. A descriptor containing parameters that can be used to estimate the complexity of the decoding process is embedded in the system stream. The encoder also encodes the video information in such a manner that the decoder can choose to ignore some of the information and only decode a subset of the encoded information in order to reduce the computational requirements. This method allows more than one decoder to decode the same bitstream giving different resolutions depending on the computational capability of the decoder.

Abstract: A system and method for clinical adaptive compression of image data is disclosed. Images are examined for contextual data, color/grayscale data, color/grayscale table, interpolated data, temporal redundancy, display mode, and/or other information. One or more subregions of data is identified, simplified, and compressed without degrading the quality of the clinically important information. The less clinically important information is more highly compressed. Redundant and non-essential data may be decimated, and the clinically necessary information is coded and made available for access.

Abstract: An image encoding device equipped with a transformation component which transforms given image data into transformation coefficients by subjecting this image data to a frequency decomposition, a dividing component which divides the transformation coefficients produced by the transformation component into a selected region on the image and a non-selected region other than the selected region, and an encoding component which encodes the transformation coefficients by preferentially allocating a greater quantity of information to the selected region than to the non-selected region, and the dividing component performs an equation evaluation of numerical equation data that stipulates the boundary of the selected region, and determines whether the transformation coefficients belong to the selected region on the basis of the results of this equation evaluation, to perform the division into a selected region and non-selected region during image encoding with less processing.

Abstract: The invention disclosed here is a method for an encoder to encode audiovisual information for transmission to the decoder without any prior knowledge of the computational capabilities of the decoder. A descriptor containing parameters that can be used to estimate the complexity of the decoding process is embedded in the system stream. The encoder also encodes the video information in such a manner that the decoder can choose to ignore some of the information and only decode a subset of the encoded information in order to reduce the computational requirements. This method allows more than one decoder to decode the same bitstream giving different resolutions depending on the computational capability of the decoder.

Abstract: An image processing method wherein a “peer group” is identified for each pixel in the image and the peer group is then used to make processing decisions at that pixel. A diffusion form of peer group processing is presented that averages over the peer group (PGA). Two parameters provide direct control over which objects are selectively enhanced: area (number of pixels in the object) and window diameter (window size needed to enclose the object). The direct correspondence between these parameters and the object characteristics is in sharp contrast to parameter selection for other image processing methods such as variational approaches in which it is unclear how to select good values for the smoothness and approximation weights. The scale space decomposition of an image determined by using the peer group number as the scale parameter leads to top hat procedures for recovering objects in a given size range.

Abstract: A decoder decodes encoded binary image data on each block in order to recover a target block of the binary image. The image decoder includes motion compensated blocking which obtains a reference block from a reference binary image by applying motion compensation using motion information. A selector selects a statistical model from among a plurality of statistical models, based on states of pixels surrounding a reference pixel in the reference block, in which the reference pixel also corresponds to a target pixel in the target block. An arithmetic decoder recovers the target block by decoding the encoded data using the selected statistical model. In one embodiment, the pixels surrounding the reference pixel in the reference block are pixels positioned within one pixel distance from the reference pixel.

Abstract: A method for processing image data includes quantizing a region in a frame with an initial quantizer level. It is determined whether an amount of bits required for encoding the region after quantizing the region with the initial quantizer level is within a bit allocation budget. The region is re-quantized if the amount of bits is not within the bit allocation budget.

Abstract: A method determines a surface of an object in a sequence of images. The method begins by estimating a boundary of the object in each image of the sequence using motion information of adjacent images of the sequence. Then, portions of each image of the sequence are ordered to produce an ordered sequence of images. The ordered portions are exterior to the estimated object boundary. Edges in each ordered image are filtered using the motion information, and each ordered image of the sequence is searched to locate the filtered edges to form a new boundary outside the estimated boundary. The filtering and searching are repeated, while projecting the new object boundaries over the sequence of images, until the new object boundaries converges to a surface of the object.

Abstract: A client-server system and method that enables efficient, low bit rate transmission of image data over a network from an image server (e.g., active cameras) to a client for, e.g., distributed surveillance. A detection and tracking module detects a human presence within an observation area and provides 2-dimensional face coordinates and its estimated scale to a video transmission module. The captured video is then efficiently encoded in log-polar coordinates using an adaptive log-polar mapping, with a foveation point centered on the face. A fovea region, which covers the target object (face), is uniformly sampled and transmitted at full resolution to the client. The periphery region (background) is sampled according to a log-polar grid.

Abstract: A method and apparatus for generating a bounding rectangle of a VOP for interlaced scan type video signals, which are capable of accurately generating a bounding rectangle, thereby preventing a color bleeding from occurring in the field-based chrominance shape subsampling. The method includes bounding rectangle formation step of forming a bounding rectangle of a video object plane in such a fashion that the bounding rectangle contains an object therein, based on input luminance shape information, optimum setting determination step of determining whether or not a spatial reference point of the bounding rectangle is positioned on an x-y coordinate of coordinate value (2m, 4n) (m and n=0, 1, 2, 3, 4, . . .

Abstract: A shape adaptive wavelet transform is performed on a digitized image signal representing an image. First, a shape information is obtained about the image, and the shape information is applied to a shape adaptive wavelet filter. The shape adaptive wavelet filter detects a starting point and an ending point of the image. The digitized image signal is divided into a low-frequency component and a high-frequency component. Then each frequency component is sampled, using the shape information. The shape adaptive wavelet filter then generates converted shape information indicating the shape for each sampled frequency component.

Abstract: When a region obtained by predetermined region division processing is a character image, the image is converted at one of predetermined compression ratios (S505). Data in a certain information amount that the region has as pixel data after image conversion is subjected to character recognition, and the degree of similarity (score) to character types registered in advance is calculated (S506). If the calculated score is equal to or smaller than a predetermined threshold value, the information amount is smaller than a minimum information amount necessary for reconstruction of the character image contained in the region. The character image is converted at a compression ratio lower than that in S505 by one step (S507-S509).

Abstract: An apparatus for coding a binary image representing an object shape includes an inferior symbol detecting unit which decides which one of binary zero and binary one is an inferior symbol that is of smaller occurrence within a given area of the binary image, a divided portion generating unit which divides a rectangular block of the given area into divided portions, a map information generating unit which generates map information for each one of the divided portions, the map information indicating whether a corresponding one of the divided portions has the inferior symbol included therein, and a coding unit which encodes only the divided portions that have the inferior symbol included therein, wherein an identification of the inferior symbol, the map information, and the encoded divided portions are output from the apparatus.

Abstract: A plurality of layers having other resolution are coded in order to encode scalably shape information. A base layer having the lower resolution is coded and transmitted to a decoder, and an enhancement layer is coded by employing a scan interleaving method from the base layer. A scan order of vertical and horizontal scannings is decided according to a generation frequency of TSD (transitional sample data) and ESD (exceptional sample data) and the number of coding bits on the base layer, or a type of boundary lines on images. When the scan order is decided, additional information indicating the scan order is transmitted to the decoder. The base layer is also coded and transmitted to the decoder. In case that the scan order is decided according to the horizontal and vertical of the boundary lines on an image of the base layer, it is no need to transmit the additional information indicating the scan order on the received base layer and execute the decoding in the same order as the coded scan order.

Abstract: An iterated image transformation and decoding apparatus and method, and a recording medium are provided. The iterated image transformation and decoding apparatus includes two polygon information generation devices for unscrambling an input coded bit stream, inputting aspect-ratio information, and reconstructing information for generating two different polygonal images; an image transformation and generation device for mapping-transforming the pixel value of an image within one of the polygons and the position of the polygon; an image memory device for storing the transformed polygonal image; and a control device for performing control so that the mapping transformation and generation of the polygon is iteratively processed. This makes it possible to provide a decoded image having a predetermined aspect ratio without a filtering process or the like which is required conventionally.

Abstract: It is an object to encode an image signal or a shape signal more efficiently than the prior art. As a means to accomplish the object, the change pixel detector 2 receives the input signal 1 as an input signal and detects the pixel which changes the two-valued pixel value. Further, the change pixel predictor 4 also reads out the reference image stored in the memory 3, and predicts the change pixel of the particular input signal. The difference value calculator 5 subtracts the output of the change pixel predictor 4 from the output of the change pixel detector 2. The difference value rounder 7 compares the tolerance value e and the prediction error D, and outputs x which requires the minimum bit number for being encoded in the value D−e≦x≦D+e. The output of the difference value rounder 7 is encoded by the decoder 8 to become the encoded signal 9.

Abstract: In a reduction process, data in portion buffers are selectively sampled at different sampling rates proportional to their assigned weights. For instance, portions assigned highest weights could be sampled at a reference rate corresponding to the repetition frequency of the originally received frames, portions assigned lowest weights could be sampled at {fraction (1/10)}th the reference rate, and portions assigned weights intermediate the lowest and highest ones could be sampled at rates less than the reference rate but more than {fraction (1/10)}th the reference rate. Accordingly, sampled portions assigned less than highest weights, but containing data representing objects in motion, could be subject to reproduction with less clarity than sampled portions assigned highest weight.

Abstract: A method for expediting a search of a digital image database comprises, for each digital image, constructing a set of at least two abstracted images varying in level of detail and searching the abstracted images. Selected ones of the abstracted images are downloaded to a user thereby reducing download time through a finite bandwidth communicative medium relative to the time needed to download complete digital images.

Abstract: A method and system can automatically select the optimal size of a visual object that can be displayed on an display device. For example, the present invention can determine the best or optimum size to display a visual object received from a computer network to a computer screen such that image degradations due to compression or transmission defects are less noticeable or perceivable. Typically, the best or optimum size to display a visual object can comprise a size that is also the maximum in which the visual object can be displayed without any degradations or defects.

Abstract: An interpolation method in which a threshold value used for determining a pixel value of a pixel generated by interpolation according to a context (state value of adjacent pixels). In the interpolation method, the ambiguity between the interpolation value and the threshold value is removed by using the context (state value of the reference pixels around the interpolated pixel), thereby reducing the blocking and smoothing phenomena in the restored binary image.

Abstract: An improved interpolation method in which a threshold value used for determining a pixel value of a pixel generated by interpolation according to a context (state value of adjacent pixels). In the interpolation method, the ambiguity between the interpolation value and the threshold value is removed by using the context (state value of the reference pixels around the interpolated pixel), thereby reducing the blocking and smoothing phenomena in the restored binary image.