Abstract: In a two-dimensional transform coding system, the transform coefficients are ordered according to the variance of the coefficients through a frame of coefficient blocks. Quantization bits are individually assigned to maximize the reduction in quantization error with each assignment. The coefficients are grouped according to the number of quantization bits assigned thereto. To assign each bit, the reduction in quantization error of the frame of the block is computed for the coefficient having the largest variance of each group as the product of the square of the variance and the normalized change in quantization error. The normalized change in quantization error may be stored in a lookup table as a function of the type of distribution of the coefficients.
Abstract: The image of a word is taken and the two-dimensional discrete Fourier transform of the image is computed. The transformed image is filtered to the first three harmonics, with both real and imaginary components. These components then make up a total of 49 unique vectors which defines a 49 orthogonal vector space. The vector space is normalized to unity and each image of a word or phrase defines a point within this 49 orthogonal, hypersphere. The same process is done to the image for the Fourier components, where there is only 25 unique vector components. Similar looking words cluster in the hypersphere and the smaller distance from one point to another defines the probability of incorrectly recognizing a word. In a study for the case of two through eleven letters in a word using both 49 and 25 vector space calculations, the results show two through eleven words are recognizable using 49 vector space and possibly the 25 vector space.
May 28, 1986
Date of Patent:
August 16, 1988
The United States of America as represented by the Secretary of the Air Force
Abstract: An automatic accurate alignment system for positioning an object to be worked which has a certain pattern on its surface at a required position. The system performs primary positioning of the object to be worked on the basis of low magnification pattern matching with respect to at least a part of the image of the object magnified at a relatively low magnification. Thereafter, the system performs secondary positioning of the object to be worked on the basis of high magnification pattern matching with respect to at least a part of the image of the object magnified at a relatively high magnification.
Abstract: A system for sensing a characteristic of a document, e.g. translucency, by using radiation from the document at a specific window area. The identified window area is illuminated to provide radiation from the area which is sensed by pixel dissection so that pixels of central location in the window area are weighted for greater significance. The pixel weighting attributes greater significance to the central portion of the defined area thereby increasing the tolerance of the system to misalignment of the window area. The observed characteristic is reduced to signal representations which are compared with registered values as to authenticate the document. Pixel-weighting techniques include overlapped pixels, filtered pixels and algebraically weighted pixels.
Abstract: An apparatus (32) and technique for compensating for motion in an electro-optical camera system. An image system is maintained on an aircraft for generating an image of terrain while the aircraft is in flight. A high resolution linear array (24) generates an image from a field of view (30), while a rectangular array (26) generates a corresponding image on a periodic basis. The rectangular image serves as a reference for the pixels of the linear array for determining if features in the linear array have shifted or distorted. An error is generated corresponding to the differences between the linear array and the rectangular array, the linear array being corrected as a function of the error.
Abstract: A method and apparatus for digitally analyzing continuous visual images, particularly with reference to the detection of mammalian cell mitotic events is disclosed. The visual images are analyzed by first extracting high frequency picture components, threshold comparison of such components and probing for annular objects indicative of putative mitotic cells. The detection of annulae is performed by an algorithm for recognizing rings of differential radii and compensating for other variations. Thereafter, spatial and temporal relationships between such objects is stored and compared to determine whether cell division occurred.
September 10, 1985
Date of Patent:
February 9, 1988
Beckman Research Institute, City of Hope
Abstract: An optical and electronic scan produces an image composed of binary value pixels. An automatic picture compression routine reduces the number of pixels in the image. A first stage four to one reduction in the number of pixels is achieved by treating each mutually exclusive four pixel unit as a separate set. A binary value is assigned to each four pixel set. That binary value is assigned to a single pixel in a first output image. The same process is repeated a second time to provide a final output image that has one-sixteenth the number of pixels as has the original input image. The value assigned to each four pixel unit set is a weighted function of the binary value of each of the sixteen pixels in a four pixel by four pixel subfield in which the unit set is centered. The greatest weight is give to the center pixels, that is, to the four pixels of the unit set. Lesser weight is given to the peripheral pixels.
Abstract: This application discloses a method and apparatus for ascertaining the fitness of a document having some printed regions and some unprinted regions, by determining its degree of soiling. A sensor (16) provides reflectance signals representing the intensity of light reflected from a number of sensed elements on the face of the document. A signal processing circuit responsive to the reflectance signals includes means (26) for determining for each of a number of areas of the document, the number of elements in the area which result in a signal level exceeding a first threshold which represents the upper limit of brightness for printed areas on the document; and means (24,34,36) for determining whether the number of elements exceeding the first threshold reaches a given proportion of the total number of sensed elements for that area.
Abstract: A pattern outline tracking system provides the functions that include scanning an object to be recognized and providing a series of analog video signals, comparing a video signal level with the specific, variable threshold level and providing a binary signal consisting of a sequence of serial bits, holding the binary signal bits as the input image pattern in an image memory, deriving partial patterns sequentially from the serial binary signal bits, each of the partial patterns consisting of a matrix of vertical and horizontal lines each containing specific numbers of pixels, matching the partial patterns against the corresponding reference patterns containing the previously defined starting points to be tracked, holding the address in the image memory for the certain pixel within the partial pattern identified in the matching step as the possible point address for the tracking operation, and tracking the outline of the input image pattern on the image memory in accordance with the possible point address.
Abstract: A fingerprint input device for reading images of a fingerprint from a surface of a finger includes a base plate, an optical scanner, a rotary member, a first cleaning member, a switch, a detector and a control member. The optical member scans the finger-impressing surface to read the fingerprint. The rotary member has a first portion and a second, window portion. The first cleaning member is provided at the bottom of the first portion of the rotary member for wiping the finger-impressing surface by rotation thereof. The switch means is operated when the fingerprint input device is to be utilized, and the detector is operative for detecting release of the finger from the finger-impressing surface. The controller is responsive to outputs of the switch and the detector for controlling rotation of the rotary member to both expose the finger-impressing surface and to clean and cover same.
Abstract: A plurality of light emitting elements which are adapted to emit light of three primary colors are activated in a sequential order, and the emitted light is transmitted through a single emission side optical fiber to illuminate an image surface. Reflected light from the image surface is conveyed to a light receiving element through a corresponding receiving side optical fiber. A decision circuit determines whether or not an output from the light receiving element has reached a given level. The sequence of emission from the respective light emitting elements as well as the amount of light emitted therefrom are controlled by control means. When the given level is reached, the associated data is stored to assure a proper reading of the image.
Abstract: An optical reading device provided with an exclusive display section for displaying character patterns and the like in response to electrical signals applied from deflection control circuits. The deflection control circuits include a scanning pattern memory for storing those data which correspond to the scanning patterns of the deflected light, and a character pattern memory for storing those data which correspond to character patterns of the deflected light, wherein a character pattern is displayed in the display section according to a scanning pattern based on the recognition of data detected by a photoelectric converter.
Abstract: Edge maps (40) derived from images are used to compute edge spectra (44), an edge spectrum having a plurality of components (41, 43) corresponding to the angular bins (60) of edge vectors having equal angular widths. Various feature detectors (56) process the edge spectrum to yield information identifying the image. A linear detector correlates a shifted prototype edge spectrum (45) to an input spectrum (44). Nonlinear detectors analyze edge spectra to detect mutually orthogonal edges and edge reversal features (90). Higher level logic (30) is used to select certain detected edge reversal features (90) as the ends of an object (16) depicted in the image.
Abstract: Circumnavigation using an increased test cell domain may be used to collect data from discontinuous characters stored in a bit map such as those printed by a dot matrix printer. The test cell domain is increased to locate a pixel and consider it a portion of the same character if it is within the minimum permissable pixel gap range. The sequence of testing must be designed to read each possible pixel and not enter an endless loop during the circumnavigation. One sequence of testing is to begin testing those cells in the column adjacent to the reference cell beginning of the cell in the row adjacent to the reference cell and testing cells adjacent to the most recently tested cell along the entire column until the desired number of cells in that column according to the minimum permissable pixel gap range have been tested. Each column adjacent to the most recently tested is tested in the described sequence until the desired number of columns have been tested to permit the minimum permissable gap spacing.
Abstract: A two-dimensional coordinate measuring pad for use, for example, to recognize handwriting. A high resistivity sheet is bounded by a lower resistivity border connected to a switched voltage source to cause a current to flow first in one direction and then in a perpendicular direction through the resistive sheet. Potential measurements at an array of precisely predetermined points are employed to generate a set of correction values which are stored in a programmable read-only memory. Thereafter, to indicate the position of an arbitrarily selected contact point on the sheet, the stored correction values for the nearest points are fetched form the memory and combined with the measured values to produce adjusted output values which more accurately indicate the positional coordinates of the contact point.
Abstract: A method and system for generating additional template images containing an object of interest in a plurality of defined orientations from a single stored digitized pixel matrix template image frame is disclosed by which the object of interest may be translated and rotated within the template frame and thereby presented in a variety of predetermined positions.
Abstract: Apparatus for defining a bank of sequential windows from a raster-scanned image of a document for storage in a data base. In addition, the apparatus includes a look-up table for each window wherein the bit pattern of the window is used to determine the correction factor which is applied to the next window as the "new" bit from a set of patterns and correction factors derived from well known algorithms developed to minimize the amount of data needed to electronically recreate a given document. The apparatus also includes storage devices to permit the application of the algorithm in overlapping pieces over the face of the document.
Abstract: A method and apparatus for removing undesirable dots and voids which are smaller than the user defined smallest data item of a bit-map image of a picture being scanned by a picture coding system. The bit-map is delayed as it is being generated to form a series of tessellations or windows of data of selected sizes. The windows are propagated through a series of neighborhood-logic elements which with the output data level of the outer ring of neighborhood-logic elements which define the window being examined to determine if those data levels are all of the same sense. If they are all of the same sense, then all of the neighborhood-logic elements which define the interior of the window are set or cleared so that the output data levels of all of the neighborhood-logic elements which define the window are of the same sense.