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: Run length encoded data derived from a raster scanner at an arbitrary angle to the horizontal direction of a rectilinear symbol such as a two-dimensional PDF code is examined to detect the skew angle between the horizontal direction of the symbol and the horizontal direction of the raster. Using the detected skew angle, a computer generates a set of virtual scan lines overlaid on the real scan lines. The computer partitions the virtual scan lines and the real scan lines into associated real and virtual line segments having common intersection points. The locations of mark to space or space to mark transitions are mapped from the real line segments onto the virtual line segments. The system can handle run length encoded data with low computational overhead.

Abstract: The present invention presents an image processing apparatus that performs rotation, enlargement, reduction, clipping or overlapping processing of images at high speed using low capacity memories instead of page memories and without the need to read image data in repetition. A first band data storing element receives input image data line by line and stores it as local data. The local data stored in the first band data storing element is transformed by local data transforming element and then stored in a transformed data storing element. The local data is then read in the order it is to be output and stored in a second band data storing element. An image output element then outputs the local data as an output image.

Abstract: A system and method for parallel computation of the unordered Hadamard transform. The computing system includes a plurality of interconnected processors and corresponding local memories. An input signal x is received, partitioned into M1 sub-vectors xi of length M2, and distributed to the local memories. Each processor computer a Hadamard transform (order M2) on the sub-vectors in its local memory (in parallel), generating M1 result sub-vectors ti of length M2, which compose a vector t of length M1×M2. A stride permutation (stride M2) is performed on t generating vector u. Each processor computes a Hadamard transform (order M1) on the sub-vectors uj in its local memory (in parallel), generating M1 result sub-vectors vj of length M2, which compose a vector v of length M2×M1. A stride permutation is performed on v (stride M1) generating result vector w, which is the Hadamard transform of the input signal x.

Abstract: A data conversion processing circuit which effects processing for data conversion which projects two-dimensional data on a plurality of straight lines with different inclinations. The data conversion processing circuit includes a circuit unit for storing two-dimensional data, a circuit unit for producing x and y coordinate locating data, which serves as an addend, according to an angle &thgr; and generating a signal for use in selecting data, a circuit unit for selecting data, which serves as an addend, from among stored data in response to the control signal, and a circuit unit for adding selected data. For producing coordinates and selecting data, the data conversion processing circuit carries out arithmetic operations for a Hough transform expressed as AH(&rgr;, &thgr;)=∫a(&rgr;·cos &thgr;−t·sin &thgr;, &rgr;·sin &thgr;+t·cos &thgr;)dt. For handling different angles &thgr;, the arithmetic operations are carried out in a time-division manner.

Abstract: A method and an apparatus of compressing data. The method and apparatus include constructing a neural network having a specific geometry using a finite and discrete Radon transform. The data is then fed through the neural network to produce a transformed data stream. The transformed data stream is thresholded. A fixed input signal is fed back through the neural network to generate a decoding calculation of an average value. The thresholded data stream is entropy encoded.

Abstract: Image data is edge-processed and then digitized. This image data is Hough-transformed to generate a parameter plane in which multiple Hough-curves are plotted in a parameter space, from which coordinates with multiple intersected Hough-curves are extracted and grouped. For each group, representative coordinates are selected to estimate slopes of linear components in the image data. In this way, the linear components can be recognized from the image data to estimate the slope, thereby making it possible to modify slopes of linear components in image data of a semiconductor wafer taken at an arbitrary angle or the like.

Abstract: A Hough transform based method of estimating N parameters a=(a1, . . . , aN) of motion of a region Y in a first image to a following image, the first and following images represented, in a first spatial resolution, by intensities at pixels having coordinates in a coordinate system, the method including: determining the total support H(Y,a) as a sum of the values of an error function for the intensities at pixels in the region Y; determining the motion parameters a that give the total support a minimum value; the determining being made in steps of an iterative process moving along a series of parameter estimates a1, a2, . . .

Abstract: There is provided an image information processing apparatus capable of detecting a defect on a device even from an unclear SEM image thereof, and without using a golden device or a CAD data, at high speed. A two dimensional Wavelet transform is applied to an input digital image data. A threshold value process (a binarization process) is then applied to longitudinal line detection components and lateral line detection components obtained by the two dimensional Wavelet transform to create respective binarization images of the longitudinal line detection components and the lateral line detection components. A Hough transform is applied to each binarization image to obtain a position and a size of an object to be detected. If a threshold value process is applied to an image in a parameter space obtained by the Hough transform, a detection of particular figure information becomes easy. With respect to the binarization image, active pixels are grouped such that the same label is given to adjacent active pixels.

Abstract: In a correlation arithmetic system adapted to detect a relative difference between two functions, an operation is simplified. This makes it possible to perform the operation with a small scale of hardware and also with great accuracy. There is adopted an operation g*h instead of the “product” in the correlation arithmetic operation. There is disclosed an arithmetic unit in which two numeral values a and b are inputted, and the two numerical values a and b are subjected to a predetermined operation process, so that a numerical value c representative of an operation result is derived. The arithmetic unit has an absolute value operation unit for evaluating an absolute value |c| of the numerical value c, and a sign operation unit for evaluating a sign “sing (c)” of the numerical value c.

Abstract: An image information recognition system is provided which quantifies an angle of rotation and/or magnification of a wiring pattern in an image to facilitate a subsequent image retrieval.

Abstract: This invention provides a method and apparatus for performing Haar transforms that process one image pixel for each clock cycle. The Haar transform is performed by cascading successive Haar transform cells where each Haar transform cell processes a 2×2 block of inputs and outputs a 2×2 block of Haar transform coefficients. Four inputs are processed by a first stage of two adders and two subtractors to generate four Haar transform coefficients. The output of the Haar transform may be quantized, run length encoded and then compressed to achieve a high compression ratio of greater than 20 to 1 for near lossless image compression.

Abstract: The image compression and expansion device comprises a recording medium, in which compressed image data, which are obtained by applying an Hadamard's transformation, a quantization and an encoding to original image data, and a quantization table are recorded. The compressed image data are decoded by a decoding unit. The decoded image data are dequantized by the quantization table. The dequantized image data are converted to corrected dequantized image data, which are then subjected to an inverse Hadamard's transformation, so that reproduced image data are generated. The quantization table is composed of quantization coefficients, each of which is to the power of 2.

Abstract: A compression/decompression engine is disclosed for reducing memory requirements of a decode system by storing decoded video data in compressed form. The compression engine comprises parsing chrominance UV data into separate chrominance U data and chrominance V data, and transform logic implementing a Hadamard transformation of multiple bytes of decoded video data in parallel into frequency domain signals. Compression logic is coupled to the transform logic and performs, preferably, a 2:1 transformation of the frequency domain signals to produce compressed video signals for storage in memory. The transform logic and compression logic transform and compress multiple bytes of decoded video data in parallel within a single clock cycle of the decode system. Upon retrieval from memory, the compressed data is returned to original format by the decompression engine, which employs the same transform logic as used by the compression engine.

Abstract: An image scanner having an improved scanning quality is disclosed. Two different color blocks having an interface therebetween are provided on the document platform of the image scanner. The image pickup device passes by the color blocks prior to the document area so that color calibrations and/or error detection can be performed before the scanning operation is performed. By this way, the color data such as the brightness of the obtained image can be compensated, and the deflection rate and/or the amplification error of the image pickup device can be corrected.

Abstract: A method for modeling a roadway and a method for recognizing lane markers based on the modeling method. The method for recognizing lane markers of roadway for a vehicle by getting image information about the roadway and information about the speed and steering angle using a camera and a sensor attached to the vehicle, includes the steps of modeling the actual roadway on which the vehicle travels, as a structure having a plurality of rectangular plates linked to each other. The modeled plates are overlayed onto the image information about the actual roadway, photographed by the camera, pixels forming the lane markers are extracted, and linear lane marker information is then obtained from the pixels forming the land markers.

Abstract: In the image pickup phase (A), right and left images are taken in through two image-pickup devices (S101, S102). Then, in the next feature extraction phase (B), right and left images are respectively subjected to feature extraction (S103, S104). Thereafter, in the succeeding matching phase (C), the extracted features of right and left images are compared to check how they match with each other (step S105). More specifically, in the matching phase (C), a one-dimensional window is set, this one-dimensional window is shifted along the left image in accordance with a predetermined scanning rule so as to successively set overlapped one-dimensional windows, and a matching operation is performed by comparing the image features within one window and corresponding image features on the right image.

Abstract: A straight-line detecting method for detecting a straight line in an image by using a combinatorial Hough transform at a high speed by restricting a voting area to assure detection of a short line segment without decreasing a processing rate, which method comprises the steps of: determining parameters of Hough transform of a straight line passing a remarkable edge point (x1, y1) and another edge point (x2, y2) in a X-Y plane according to the transform equations .theta.=-a tan{(x1-x2)/(y1-y2)} and .rho.=x1 cos .theta.+y1 sin .theta.; voting of pixels corresponding to parameters (.theta., .rho.); and detecting a straight line existing in the X-Y plane by the number of votes in a plane of .rho.-.theta. parameters and which is characterized in that another edge point (x2, y2) is a point existing in any of specified neighboring areas positioned at right and left and above and below from a center area where the remarkable edge point (x1, y1) is located.

Abstract: The invention relates to a method and an apparatus for the analysis and correction of colour casts in a digitally represented image. According to a preferred embodiment, the crominance contents of the image may be transformed by use of a Hough-transformation or the like transformation, wherein the line structures in the carthesic system of co-ordinates for the crominance content is transformed to a representation in a Hough-diagram in the form of a polar system of co-ordinates, since such line structure is hereby transformed into a point structure. The Hough-diagram is examined with a view to finding a cell in the diagram which has the highest value and the cell found is saved. subsequently cells are neutralized in the Hough-diagram within a predetermined angular distance, and the Hough-diagram is re-examined with a view to finding that diagram cell which, following neutralization, has the highest value.

Abstract: In a correlation arithmetic system adapted to detect a relative difference between two functions, an operation is simplified. This makes it possible to perform the operation with a small scale of hardware and also with great accuracy. There is adopted an operation g*h instead of the "product" in the correlation arithmetic operation. There is disclosed an arithmetic unit in which two numeral values a and b are inputted, and the two numerical values a and b are subjected to a predetermined operation process, so that a numerical value c representative of an operation result is derived. The arithmetic unit has an absolute value operation unit for evaluating an absolute value .vertline.c.vertline. of the numerical value c, and a sign operation unit for evaluating a sign "sing (c)" of the numerical value c.

Abstract: A hadamard code generation circuit is disclosed. The circuit includes a start reset signal generator for generating a start reset signal START.sub.-- RESET when a 6-bit output signal REF.sub.-- C from the 6-bit reference counter, a higher 4-bit index output signal H(5:2) of the 6-bit register and a 2-bit value from a ground circuit are identical; a "0" value force allocation unit for outputting a FORCE.sub.-- 0.sub.-- DEL signal for forcibly allocating all values of the 0th column to "0" when a 4-bit output signal REF.sub.-- C (5:2) from the 6-bit reference counter and a 4-bit value from the ground circuit are identical; a 2-bit counter for receiving the start reset signal and an external clock signal, outputting lowest bit signals C1 and C0 and outputting a carry-out signal; a 4-bit counter operated in accordance with a result that an inverted FORCE.sub.

Abstract: A method and apparatus for the automatic segmentation of the pectoral muscle boundary in a digital mammogram is disclosed. Parameters of the pectoral boundary line are determined by calculating gradient magnitudes of the digital mammogram in a region of interest, accumulating the gradient magnitudes into a parameter plane according to a Hough transform, and using information in the parameter plane for identifying the pectoral boundary. Prior to being accumulated in the parameter plane, the gradient magnitudes are weighted by a monotonic, multi-bit weighting function designed to exhibit larger variations near more densely populated levels of gradient magnitudes, and designed to exhibit smaller variations near less densely populated levels of gradient magnitudes, for allowing greater robustness against variations in gradient magnitudes among different digital mammograms.

Abstract: In an Hadamard transform coefficient predictor for 8-point and/or 8.times.8-point Hadamard transform, when the transform coefficients of one-dimensional 8-point Hadamard transform are represented by y(0), y(1), . . . , y(7) from the lowest order, y(1) is multiplied by 1/2 and output as a prediction value of y(3), y(2) is multiplied by 1/4 and output as a prediction value of y(4), y(2) is multiplied by 1/2 and output as a prediction value of y(6), and y(1) is multiplied by 1/4 and output as a prediction value of y(7). Alternatively, when a and b represent real numbers which are above zero and below 1, the multiplication value of y(2) and b/2 and the multiplication value of y(4) and (2-2b) are added to each other, and the addition result is output as a prediction value of y(6). Further, the multiplication value of y(1) and a/4 and the multiplication value of y(3) and (1-a)/2 are added and the addition result is output as a prediction value of y(7).

Abstract: An encoding process is configurable to trade calculational complexity for bit rate while maintaining code format and image quality. A software encoder which implements the encoding process monitors bandwidth and processor utilization and when either is overtaxed, adjusts complexity to compensate. If further adjustment is required, encoded image quality is changed. Changing the search process or search window size used during searches for predictive blocks or changing a threshold which is compared to inter-code length before determining whether to calculate an intra-code changes complexity. An intra-coding method subtracts a base which depends on codes representing neighboring blocks from coefficients for a current block to determine a difference which is often small and easily compressed. Additionally, restricting a quadtree structure for one color component to a sub-tree of a quadtree structure for another other color component improves compression efficiency and reduces coding complexity.

Abstract: To improve an area cut-out ratio, a difference value between coefficients which are neighboring in an image block is calculated by a difference value calculation unit, an average value and a dispersion value of the difference value data are calculated by a distribution state calculation unit, and the average value and the dispersion value are compared with preset threshold values by a threshold value determination unit, thereby determining whether the relevant block is a character pattern area or not.

Abstract: A method for recognizing text in graphical drawings includes creating a binarized representation of a drawing to form an electronic image of pixels. The image is discriminated between text regions and lines in the image by grouping pixels into blocks and comparing blocks with a predetermined format to identify text regions. The lines that remain in the text regions are removed to create text only regions. The text is recognized in the text only regions.

Abstract: A method of transferring digital image data over a communication link transforms and orders the data so that, as data is received by a receiving station, a low detail version of the image is immediately generated with later transmissions of data providing progressively greater detail in this image. User instructions are accepted, limiting the ultimate resolution of the image or suspending enhancement of the image except in certain user defined regions. When a low detail image is requested followed by a request for a high detailed version of the same image, the originally transmitted data of the low resolution image is not discarded or retransmitted but used with later data to improve the originally transmitted image. Only a single copy of the transformed image need be retained by the transmitting device in order to satisfy requests for different amounts of image detail.

Abstract: For providing a digital representation of an optical scene, radiometric values of N.times.N points representing a real two-dimensional optical scene are measured in the form of N.times.N respective electric signals x(i,j). N.times.N masks Mnm(i,j) respectively associated with the N.times.N electric signals x(i,j) are generated. For each couple (n,m), the N.times.N electric signals x(i,j) are weighted by the corresponding N.times.N masks Mnm(i,j) and the weighted electric signals are added to obtain N.times.N coefficients of the Walsh-Hadamard transform representing the optical scene.

Abstract: A method of processing a digital radiographic image includes the following the steps. (1) A digital radiographic image having a radiation region and a collimation region defined by collimation blades at least partially bounding the radiation region is provided. (2) Pixels of the digital radiographic image are detected and classified as collimation boundary transition pixels. (3) Candidate collimation blades are detected and classified from the collimation boundary transition pixels. (4) At a regional level, the collimation region are determined from the results of the line-level delineating step.

Abstract: A recognizing apparatus for estimating an environment description parameter, and substantially reducing a memory capacity required to represent a parameter space. The recognizing apparatus stores a restricted parameter space as a parameter space restricted by predetermined observation point information and based on an object model as shape information. The apparatus votes for a parameter subset consistent with the object model for each characteristic point obtained through environment observation in a restricted parameter space. The apparatus then outputs an estimated value for a environment description parameter according to the result of the voting for the restricted parameter space.

Abstract: An image processing apparatus for extracting lines from an image using the Hough transform. A processor element is assigned to each quantization point in a Hough space. Each processor element calculates, for each scanning line of the image, intersections of the scanning line and the line corresponding to this processor element once per scanning line. A black pixel (a Hough transform object point) on the scanning line is also obtained sequentially. The coordinate values of the intersection are compared with those of the Hough transform object point, and when they agree, voting to a ballot box memory of the processor element is performed. The voting results become Hough transform data. This makes it possible to implement a high speed Hough transform, and to reduce the size of the apparatus.

Abstract: An image processing apparatus which eliminates influence of motion of background and performs stable text-image cutout at a high extraction rate. In the apparatus, a block divider 20 divides image data in frame units into blocks of a predetermined size. A DCT unit 24 performs orthogonal transformation on each divided block, and a harmonic coefficient calculator 25 extracts data for judging whether or not the block is a character area. A still-block detector 26 compares the data with data of a corresponding block after orthogonal transformation in a previous frame, judges whether or not the difference between these data is small or large, and outputs the result of judgment. An attribute judgment unit 23 judges whether or not the block is a character area, based on the data from the harmonic coefficient calculator 25 and the data from the still-block detector 26, and outputs the result of judgment.

Abstract: Based on left and right images that have been taken by image pickup devices, a correlation processing section obtains parallax data as three-dimensional data of objects. At processing start time, a plane estimating section estimates the position of a planar object such as a road or a floor by utilizing the Hough transform based on the parallax data that have been obtained successfully for part of rectangular segments of the images, and interpolates the parallax data by using the position of the planar object. At a time point after the processing start time, a parallax variation detecting section detects an object by comparing current parallax data with the parallax data at the processing start time.

Abstract: A method for forming plane signal (image) comprises modulating signals, applying to inputs of reproducing elements and summing current components of modulated signals.A device for forming images comprises matrix of reproducing elements, each of the element being provided with signal modulating and signal summing means.A method for transforming plane signal (image) comprises applying signals to receiving elements, modulating image signals, formed in accepting cells of receiving elements, summing in every column results of modulation, applying signals modulating so got sum signals and summing results of modulation.A device for transforming image comprises matrix of receiving elements, each of elements being provided with signal modulating mean, also column lines are provided with signal summing and signal modulating means.

Abstract: A system and method for filtering edge points in an edge image represented by an array of edge image coordinates each having an edge point magnitude. An edge map array is formed of edge map coordinates each assigned one of three edge indicator values indicative of the relationship between the corresponding edge image coordinate edge point magnitude and two preselected edge magnitude thresholds. An output edge point array is formed of output edge point coordinates to indicate those edge image coordinates and corresponding object image pixels that are validated by the system as being associated with likely object edge points. This is accomplished by constructing a first address stack to track memory addresses at which edge map coordinate values are stored and constructing a second address stack to track memory addresses at which output edge point coordinate values are stored as the output edge point array is constructed.

Abstract: A lane image processing system for a vehicle to recognize the lane on which the vehicle travels. The system has a CCD camera mounted on the vehicle for viewing a roadway scene ahead of the vehicle to output image data of the roadway scene including the lane on which the vehicle travels. Obstacles such as preceding vehicle are detected and a region for image data processing is determined so as to avoid the position of the obstacle and the lane on which the vehicle travels is recognized based on the processed result. More specifically, lane boundary positions estimated up to a preceding cycle are stored as historical information and in the region Hough transformation is limited in areas determined based on the historical information and the horizon in the roadway scene.

Abstract: A two-dimensional discrete cosine transform computing circuit has a transform block, formed by a multiplication unit and a Hadamard transform unit, which performs computations in a sequence of Hadamard transforms followed by multiplications in a two-dimensional discrete cosine transform, and in a sequence of multiplication followed by Hadamard transforms in an inverse two-dimensional discrete cosine transform. A memory block temporarily stores input/output data of the transform block; an input/output processing block performs pre-processing and post-processing such as cumulative addition and subtraction of input/output data of the transform block. A control block controls, when performing processing for either a two-dimensional discrete cosine transform or inverse two-dimensional discrete cosine transform, multiplications of coefficients in the transform block, read/write of the memory block, and the input/output processing block.

Abstract: A optical processor facilitates the alignment between an original image and transformation patterns, thus improving the operation accuracy, and, if a fixed transmission mask is to be used, facilitates the fabrication of the fixed transmission mask. The optical processor includes an original image display to display an original image and a pattern display to display the patterns for a transformation on the original image displayed on the original image display. In the optical processor, the original image display is os made as to have pixels with a smaller aperture ratio than that of the pixels of the pattern display.

Abstract: An image of a part is used to judge whether it is good or no good. A predetermined intensity, edge or similar feature image is extracted from the inspected image and is quantized. The quantized feature image is obtained from a training image in advance, and the feature is subjected to a generalized Hough transform operation which refers to the weight of each feature point for each feature value in the image and vote the feature point weight. The similarity is obtained from the results of the transform operation, and a good/no good judgement is made depending upon whether the similarity is above or below a threshold value.

Abstract: A system for detecting lines in images using line neighborhoods and a parallel coordinate transformation. The process introduces the concept of line neighborhoods to accommodate the uncertainty in line detection arising from image noise. Because line neighborhoods in Cartesian coordinates have ambiguous and unbounded regions and always overlap one another, a parallel coordinate transform is used to transform Cartesian coordinate image plane line segments to points in a bounded and nonambiguous region of the parallel coordinate transform plane. Line detection then becomes a simple problem of detecting point clusters in the parallel coordinate transform plane.

Abstract: An image line-segment extracting apparatus which in processing steps to conduct Hough transformation of a sequence of edge pixels composing an image, to plot a histogram based on Hough function values, to successively detect peaks from the histogram and to extract a plurality of line segments existing in the image, can detect the only true peaks, effectively eliminating false peaks. The process also includes a feature by which region is defined and the amount of influence of a remarkable one of the peaks in a histogram upon frequency distribution on the histogram are determined and then histogram frequency values of other peaks in the defined region are corrected according to the determined amount of influence to partially rewrite the histogram.

Abstract: A square anti-symmetric uniformly redundant array coded aperture includes transparent and opaque cells and exhibits a normal mask pattern at a first position and a complementary mask pattern when rotated to a second position rotationally offset by 90.degree. from the first position. The coded aperture is utilized in a coded aperture imaging system for imaging a source of non-focusable radiation such as a gamma ray or x-ray emitting source.

Abstract: A new type of fully-multiplexed imaging device is described for use at wavelengths where efficient focal plane array detectors are unavailable. It shares some properties in common with the familiar technique of cycle-redundancy (i.e. Hadamard transform) imaging, but many of its features and capabilities are unique. Some of these characteristics are, first, the new approach employs image encoding masks that are both transmitting and reflecting, thereby increasing optical efficiency, and second, the technique requires only 2.sqroot.N encoding masks to image a field of N pixels, a dramatically smaller number than that needed (.about.2N) by traditional methods. Dual complementary inputs are used for first-order passive rejection of radiation background interference. The resulting image spatial resolutions along two coordinate directions are completely independent of each other. The measured image is formed by a convergent tiling of the image plane, quite unlike conventional raster scanning.

Abstract: In accordance with the present invention, a method and apparatus are disclosed for real-time processing of digitized image data using an implementation of the Hough transform. More specifically, frames of image data are received and stored as pixel arrays in a rectangular coordinate system. Pixels having values above a predetermined intensity are subsequently transformed into polar coordinate space. Maximum meeting points of curves associated with all pixels which have been transformed into polar coordinate space are then identified using a real-time, hardware implementation of a neuro-Hough transform for analyzing windowed portions of the digitized image. The present invention identifies maximum meeting points in polar coordinate space using a neuro-Hough transform implemented as a single chip.

Abstract: A field recursive type noise reduction apparatus is disclosed in which the field difference is divided into a plurality of frequency components by means of the Hadamard transformation and the existence of a boundary of an oblique line in the proximity of pixels to be processed is detected from frequency components remaining by removing minimum and maximum horizontal frequency components by an oblique line detection circuit. In the location where the existence of the boundary of oblique line is detected, a feedback rate of the apparatus is reduced to avoid deterioration of the image quality by nonlinear processing means and in other locations the feedback rate is increased to improve the S/N, so that a noise reduction apparatus is realized having excellent improvement of the S/N as compared with a conventional field recursive type noise reduction apparatus.