Abstract: An image processing apparatus of this invention is arranged so that a threshold for use in binarization is switched between two types in accordance with whether or not a portion in which input pixel data is located is positioned in the vicinity of an edge portion of an image. One type is a threshold which varies in the width of variation corresponding to the density of the input pixel, while the other type is a fixed threshold. If an objective pixel is located in the vicinity of an edge portion of an image, binarization is carried out with the fixed threshold, while if the objective pixel is located in a non-edge portion, the variable threshold is used.

Abstract: During segmentation of individual characters in a column, the invention predicts for the next character an optimal minimum gray level sensitivity threshold while scanning pixel rows between characters by lowering the threshold unless or until noise in the image causes the scanner to detect too many "ON" pixels between adjacent numerals. During subsequent attempts to recognize the segmented character, the invention computes a confidence score and increases the sensitivity threshold above the predicted level if the confidence score is insufficient.

Abstract: A digital processing architecture for a high resolution image sensor uses a plurality of like digital processors for time-divided processing of the output of the sensor. Each processor is operational according to start and stop signals from a programmable sequencer. In a preferred embodiment, two sets of processors handle a line resolution of 1024 pixels, one set doing the first half of each line and the other set doing the second half. This is of particular utility where vertical processing is required, and the full line delays needed are divided into partial resettable delays resident in each of the processors.

Abstract: A technique for optical recognition of a designated character, which forms alternating dark and light bands from an analysis of as many as seven sub-regions of the character. Only the dark-and-light sequence of bands, not the relative or absolute sizes of these bands, is used to identify the character. The dark-and-light sequences found for the designated character are each compared with the corresponding sequences for a group of candidate characters available in a data base. Using a binary decision tree analysis, one or more candidate characters is identified whose sequences match the corresponding sequences for the designated character. If more than one candidate character is thus identified, any of seven additional processing techniques can be used to reduce the number of surviving candidate characters to a single character.

Abstract: Data which represents a border line of a region of a video image to be observed and which contains a large luminance difference compared with that of a defect to be detected is excluded in judgement when it belongs to a predetermined region so that it is possible to detect a minute defect or defects reliably even when a position of a workpiece is displaced from that of a master or when the border line of the workpiece is not coincident with that of the master due to a difference in size therebetween.

Abstract: An image data coding apparatus adopting the method of this invention includes a divider section for dividing an image having gray levels into a plurality of blocks, each consisting of a plurality of pixels; a pre-processing section for detecting the density and the presence/absence of an edge in an image of each divided block; and a vector quantizing section for encoding an image of each block in accordance with the detected density and the presence/absence of the edge in the image of the block. The pre-processing section detects a standard deviation of density levels of the pixels included in each block. The vector quantizing section divides the density levels of the pixels included in each block into a plurality of level sections so as to obtain a code indicating the level section in which the density level of each pixel is included. In the vector quantizing section, the obtained code is again encoded by means of vector quantization.

Abstract: Check processing apparatus including an offset pair of image-lift sites and an associated "Heat-Dump", with this apparatus preferably built as an add-on, plug-in module to be used in a check-sorting system.

Abstract: The invention provides a method and apparatus for performing image data compression. Initially, a two-dimensional block of image data in a spatial domain is transformed by a transform coder, resulting in a two-dimensional array of activity coefficients in a frequency domain. The array is then serialized and quantized, yielding a one-dimensional array of coefficients with a leading coefficient and an original trailing coefficient. Next, a portion of the array is selected by choosing a new trailing coefficient based on a setable ratio of the energy of the new short array to the energy of the original array. Lastly, an end-of-block symbol is appended after the new trailing coefficient, and the selected portion of the array is encoded using an entropy coder. The invention allows image data to enjoy enhanced compression with good image fidelity, and allows image fidelity to degrade gracefully in trade for higher degrees of image compression.

Abstract: A method of detecting a disparity in a cyclic length of a printed pattern includes the steps of scanning a to-be-checked sheet on which predetermined patterns are cyclically printed in the longitudinal direction by an electronic camera to obtain image data corresponding to a pattern, extracting outline data representing an outline of the pattern from the image data, writing the outline data extracted in the extraction step into a memory with a predetermined cycle in accordance with travel position data of the to-be-checked sheet, reading out the outline data, and comparing first outline data for one cycle read out from the memory and second outline data which is read out after the first outline data is read out. Disparity of a cyclic length of the pattern is detected based on a correlation between the first and second outline data, e.g., a coincidence or noncoincidence therebetween, in accordance with the comparison result in the comparison step.

Abstract: There is an image processing apparatus to perform an image editing process such as to extract a part of an image or to synthesize a plurality of images. This apparatus comprises: an image information generator to generate image information; an area information generator to generate area information to extract a desired area of the image information; a coding circuit to compress and code the area information; a memory to store the area information which was compressed and coded by the coding circuit; a decoder to decode the area information in the memory; and an extracting circuit to extract the desired area of the image information on the basis of the area information which was decoded by the decoder. The coding circuit codes the area information on a unit basis of n.times.m pixels by use of a coding look-up table. The decoder also decodes the area information by use of a decoding look-up table. The image information is separated into lightness information and shape information on a unit basis of n.times.

Abstract: A method and apparatus is provided for written character recognition. The method includes thresholding an input gray level character image to derive its bilevel image, extracting stroke-based features of said bilevel image, classifying the features based upon a set of predetermined rules, determining the identity of the character utilizing the classification, and indicating the identity of the character. A method and apparatus for extracting features of a written character is also provided. This method includes raster scanning the written character to determine stroke-based features, extracting vertical strokes of the character and extracting horizontal strokes of the character.

Abstract: Methods and apparatus are disclosed for rapid and interactive "warping" of a first image made up of pixels to form a resulting image made up of pixels which are aligned, pixel-for-pixel, with a second image made up of pixels. The images may be stroboscopic voltage contrast images representing operating states of two integrated circuit devices--a failing device and a fully functional device. The aligned images permit an engineer who may have little knowledge of the device to diagnose dynamic failures of the failing device by comparing the aligned images to produce an image showing the differences.

Abstract: A polygon-based graphics/text separation method is comprised of two sequential processes. First a raster to contour vector conversion step is used to convert a digitized bitmap into a collection of simple polygons. Next a component classification process is used to extract six particularly defined features of each of the individual polygon-based components to enable the separation of graphics and text polygons. Graphical polygons are further classified into four subclasses. Textual polygons are grouped into polygon strings (text strings). Each string contains a sequence of segmented character contour polygons which is ready for an optical character recognition algorithm to convert them into computer understandable ASCII characters.

Abstract: A high-speed optical fingerprint correlator. The invention uses the unknown ingerprint to produce a matched filter for autocorrelation and detecting and displaying the resultant indicia of correlation, the correlation peaks produced. The matched filter is produced by performing digital Fourier transform on the optically scanned digitized image input of the unknown print. The matched filter is then multiplied with the optically performed Fourier transforms on the reconstructed known fingerprint images provided in digital format from visual or digital storage.

Abstract: From a series of primary image signal components of an image signal representing an image, secondary image signal components, which correspond to an enlarged image or a reduced image and the number of which is different from the number of the primary image signal components, are obtained from predetermined interpolating operation formulas. Values of coefficients in the predetermined interpolating operation formulas, which values correspond to a desired sharpness of an enlarged image or a reduced image, are calculated from values of the coefficients in the predetermined interpolating operation formulas, which values are different from each other and which values correspond to enlarged images or reduced images having the same scale of enlargement or the same scale of reduction and having different sharpness. Secondary image signal components are then obtained from the predetermined interpolating operation formulas in which the coefficients are set to the values thus calculated.

Abstract: A neural network system and method that can adaptively recognize each of many pattern configurations from a set. The system learns and maintains accurate associations between signal pattern configurations and pattern classes with training from a teaching mechanism. The classifying system consists of a distributed input processor and an adaptive association processor. The input processor decomposes an input pattern into modules of localized contextual elements. These elements in turn are mapped onto pattern classes using a self-organizing associative neural scheme. The associative mapping determines which pattern class best represents the input pattern. The computation is done through gating elements that correspond to the contextual elements. Learning is achieved by modifying the gating elements from a true/false response to the computed probabilities for all classes in the set. The system is a parallel and fault tolerant process.

Abstract: An electronic filing device, comprising an electronic computer such as a personal computer and an image processing device, in which the electronic computer is provided with a data port and a video port and controls the image processing device by using commands and data given to the image processing device by the electronic computer through the data port and image information outputted by the electronic computer through the video port is enlarged within the image processing device, synthesized with image data searched by the image processing device, and displayed on a display device in the image processing device.

Abstract: A method of identifying an object from a digitized image is disclosed. The method is particularly applicable to the classification of ships. The digitized image of an unknown ship is represented as a set of vertical vectors. The vertical vectors are normalized and then compared to a library containing normalized digitized images of known ships. The unknown ship is thereby identified as being similar to the known ship corresponding to the library image most closely matching the image of the unknown ship.

Abstract: A pattern generation method for registering handwriting and a method for updating same that facilitates the obtaining of a registration pattern that reflects accurately the characteristics of an individual handwritten signature. Dynamic programming matching is used to obtain a cumulative error among a plurality of input patterns in the form of a distortion function, and this distortion function is used for shape averaging and time distortion averaging of input patterns.

Abstract: A method and an apparatus for processing the predetermined numeric fields containing the handwritten numeric entries. The numeric fields are subjected to character recognition analysis, and a confidence level is associated with each such numeric field reflecting the degree of confidence with which the apparatus has recognized the numeric dollar amounts. The transaction is automatically balanced from the data in the system by computing a combined value of the recognized dollar amounts making up a transaction and comparing the combined value with the transaction total. This computation and comparison is performed automatically, without the prior intervention or assistance of a human operator to review, verify, correct or otherwise modify the dollar amounts recognized by the system.