Abstract: A method of exposing a semiconductor substrate with an exposure apparatus and correcting the position, measured by a detector of the exposure apparatus, of the surface of the semiconductor substrate with respect to a focal plane of a projection optical system, includes the steps of printing an inspection pattern on a semiconductor inspection substrate with the exposure apparatus, developing the printed inspection pattern, obtaining an image of the developed inspection pattern with a charge-coupled device camera, calculating a measurement error in the position of the surface of the semiconductor inspection substrate measured by the detector during exposure of the semiconductor inspection substrate based upon the obtained image and storing the calculated measurement error in a controller, and correcting the measured position measured by the detector during subsequent exposure of the substrate by the exposure apparatus based upon the calculated measurement error stored in the controller.

Abstract: A method is provided for setting up a photon counting detector, to enable it to carry out a centroiding procedure in which a photon event occurring in a given range is allocated to one of a plurality of channels into which the range is divided. The method comprises the steps of: a) sub-dividing the range into a plurality of channels of preferably equal width and loading the channel boundaries into a look-table; b) performing an integration on a flat field and allocating photon events to the channels defined in step a); c) counting the number of events allocated to each channel; d) altering the channel boundaries in directions tending to equalize the number of events allocated to each; and, optionally, e) repeating steps b), c) and d) until the variation between the number of events held by the channels is at an acceptable level.

Abstract: An electronic document-imaging arrangement wherein imaging means generates imaging-bits representing a given document and transfers these bits on a "per-document basis" to various successive electronic processing stages and, finally, to a data base storage unit (DBS); this arrangement also including a tag stage for creating tag bits unique for each such imaged document; these tag bits to be transferred with the imaging bits for each document to each such processing stage that handles the imaging bits, on a per document basis, and finally to DBS interface means for final matching and removal of the tag bits.

Abstract: A method of multithresholding a document image considers both local and global information in determining the number of thresholded levels contained in a gray-scale image. A gray-scale image is comprised of a plurality of pixels. Each pixel has an intensity value. Local regions of pixels having similar intensity values are identified and compiled into a global representation of the image. Intensity ranges are identified in the global representation in which the number of regions of pixels are approximately constant. An intensity value is identified within each range which represents the most constant point within the representation range. The number of threshold levels is set equal to the number of ranges in the representation.

Abstract: Groups of symbols to be recognized are standardized by fitting four flexible curves to the group of symbols. The curves are fitted by minimizing a cost or energy function that associates a cost with the curvature of the curves, the slant of the curves, the displacement in spacing between the curves and the distance of maxima and minima points from the curves. After the curves are fitted to the group of symbols, the symbols are standardized by transforming coordinates systems so that the fitted curves are placed in a predetermined configuration.

Abstract: An image processing device which comprises an input terminal 1, 100 for inputting an image signal, a dividing circuit 29, 101 for dividing the image signal into a plurality of code groups, and a plurality of processing circuits 31a, 31b, 31c, 31d, 105, 106, 107 for processing the plurality of code groups respectively. Each of said plurality of processing circuits includes a motion detection circuit 21, 206, 207 for detecting a motion of the image, and a handling circuit, 111, 112, 113, 114 for handling a part of the code group to another processing circuit.

Abstract: The apparatus for the recognition of speech comprises an acoustic preprocessor, a visual preprocessor, and a speech classifier that operates the acoustic and visual preprocessed data. The acoustic preprocessor comprises a log mel spectrum analyzer that produces an equal mel bandwidth log power spectrum. The visual processor detects the motion of a set of fiducial markers on the speaker's face and extracts a set of normalized distance vectors describing lip and mouth movement. The speech classifier uses a multilevel time-delay neural network operating on the preprocessed acoustic and visual data to form an output probability distribution that indicates the probability of each candidate utterance having been spoken, based on the acoustic and visual data.

Abstract: A method for imaging or reconstructing an image of an object includes the following steps: obtaining multiple original views of the object, each original view being obtained at a different relative orientation of the object from each other view; combining the multiple original views in an image plane to obtain a first composite image; phase inverting the first composite image to obtain a phase inverted composite image; combining the phase inverted composite image with each original view to obtain a second composite image; determining a differential between the second composite image and each original view to obtain a plurality of differentials; applying the plurality of differentials to the second composite image to obtain a resulting composite image; and optionally amplifying the resulting composite image to obtain a reconstructed image. Apparatus for carrying out the method may use analog circuitry, a high speed computer or a combination of analog and digital techniques may be used.

Abstract: A data compression system decorrelates input data into sets of frequency component data. Different sets of frequency component data are entropy encoded using different coding tables matched to the data values characteristic of the sets of frequency component data they are used to compress. Matching the coding tables to the sets of frequency component data improves compression. The order in which the data values from within each set of frequency component data is fed to the entropy encoder is varied to exploit any intra-set correlation.

Abstract: The feature points of the input video signal are detected by canny edge detecting method, the coordinates of the feature points detected are detected is chain coded, the length of the region, including the feature points, of which edge intensity is the specified value or more in the edge intensity maximum change direction at the feature point coordinates obtained when the feature point is detected, are coded. The chain-coded data and the region length coded data are then multiplexed. Thereby, the amount of bits required for coding of the coordinates information can be reduced, as a result the total amount of codes can be reduced.

Abstract: A reference-character size determination unit determines a reference-character size based on a distribution of sizes of particular continuous image parts, the continuous image parts constituting an image. An image part sorting unit sorts the particular continuous image parts into character image parts and other image parts based on the reference-character size. A rectangle determination unit determines a rectangle for each continuous image part, a size of the rectangle corresponding to a size of the corresponding continuous image part. The reference-character size determination means determines the reference-character size based on a distribution of sizes of the rectangles. The image part sorting means sorts the particular rectangles into rectangles for the character image parts and rectangles for the other image parts based on the reference-character size.

Abstract: An image sensor has a light receiving section having a number of outputs along one dimension that is the same as the number of inputs along a dimension of a block serving as a unit for image processing, thereby eliminating the need for a buffer memory.

Abstract: An image signal and a series of coordinates representing a contour of region of the image signal are input. In the first step, the region contour is approximated by a polygonal shape. In the second step, bits at addresses in a bit map memory corresponding to pixels within the approximated polygonal shape are made on. In the third step, coordinates indicative of apexes of the approximated polygonal shape are coded as first series of codes. In the fourth step, only image signals which are indicated by the bit map memory as existing within the approximated polygonal shape are coded as a second series of codes. In the fifth step, the first and second series of codes obtained in the third and fourth steps are multiplexed and output.

Abstract: A system for selecting decorative materials is based on large numbers of high-resolution, full color images of decorative materials stored in a compressed format on an inexpensive medium such as a CD-ROM. In creating the ROM, each image is coordinated with additional information such as style of pattern, type of material, and other auxiliary information. Before final compression and storage of the image data color information is added by spectrophotometrically analyzing the decorative material. Color values for a background color and up to four foreground colors are determined. Individual colors are then referenced to a comprehensive color standard system containing a large number of standardized color swatches. Spectrophotometric color referencing allows the data records to be rapidly searched on the basis of color, as well as the other information in the record. Wallpaper patterns, drapery material, floor covering, or paint can then be rapidly selected on the basis of matching color.

Abstract: A method for use in an object-oriented analysis-synthesis coder, for segmenting an image of a current frame into moving objects and describing each motion of the moving objects with a set of motion parameters, which comprises the steps of: (a) comparing the current frame with a preceding frame to detect a motion vector field; (b) sliding window blocks on the motion vector field, and computing the degree of homogeneity for each of the window blocks to produce a seed block; (c) determining a set of initial values for the set of motion parameters based on the seed block; (d) evaluating whether each motion vector contained in and around the seed block is describable with the set of initial values or not, to thereby detect a revised region; (e) determining a set of revised values for the set of motion parameters from the revised region; (f) evaluating whether each motion vector contained in and around the revised region is describable with the set of revised values or not, to thereby detect a newly revised region;

Abstract: A system for locating at least one object within an intensity image is disclosed. The system includes means for analyzing the intensity image with at least two different edge detectors to detect a multiplicity of edges oriented in at least two directional ranges. First comparing means is provided for comparing information representative of detected edges angularly oriented in a first directional range and information representative of detected edges angularly oriented in a second directional range. Means for selecting at least one foreground region in the intensity image is also included. The detected edges in the selected foreground region are angularly oriented substantially in accordance with a target orientation.

Abstract: An imaging device for imaging an object is situated above the object. A downward-illumination device for emitting light to the object is situated as one unit with the imaging device. A background-illumination device for emitting light to the object is situated below the object. An image obtained by the imaging device is stored in a memory unit of an image processing apparatus. The image processing apparatus includes a processing unit for logically processing the image. The illumination devices are connected to power supplies. The image processing apparatus and the power supplies are controlled by a controller. Thus, a bonding position is recognized easily with simple structure, the reliability of test results is enhanced, and the cost for testing is reduced by a decrease in the processing time.

Abstract: A neural network is used to recognize characters from a character set. Based upon the character recognized, a smaller neural network is used for verification of the character recognized. The smaller neural network is trained to recognize only a single character of the set and provides a "yes" or "no" type verification of correct identification of the character.

Abstract: In brief, a method of reducing an M X N input binary image (M rows of N pixels each) by a factor of m vertically and n horizontally includes the steps of performing at least one logical operation between bits in consecutive groups of m adjacent rows to provide a resultant single row for each group of m rows, and performing at least one logical operation between bits in consecutive groups of n adjacent columns to provide a resultant single column for each groups of n columns. For certain types of reductions, the resulting reduced image will be the desired output image, while for other types, the resultant image will be one of a required plurality of intermediate images, which are then combined to provide the desired output image.

Abstract: A method, for use in an object-oriented analysis-synthesis coder, for segmenting an image signal into moving objects contained in a current frame and describing each motion of the moving objects with a set of motion parameters using a multi-stage segmentation comprises the steps of: (a) comparing the current frame with its preceding frame to detect a moving area in the current frame; (b) further comparing the current frame with the preceding frame to detect a motion vector for each pixel in the moving area, i.e., a motion vector field; (c) first segmenting the motion vector field by using a two dimensional translational patch model; (d) secondly segmenting the first stage segmenting results by using a planar patch model; and (e) thirdly segmenting the second-stage segmenting results by using a parabolic patch model.