Abstract: A method for constructing a data structure for a data string of characters includes producing a matrix of sorted rotations of the data string. This matrix defines an A array which is a sorted list of the characters in the data string, a B array which is a permutation of the data string, and a correspondence array C which contains correspondence entries linking the characters in the A array to the same characters in the B array. A reduced A' array is computed to identify each unique character in the A array and a reduced C' array is computed to contain every s.sup.th entry of the C array. The B array is segmented into blocks of size s. During a search, the A' and C' arrays are used to index the B array to reconstruct any desired row from the matrix of rotations. Through this representation, the matrix of rotations can thus be used as a conventional sorted list for pattern matching or information retrieval applications. A data structure containing only the A', B, and C' has very little memory overhead.

Abstract: An environment recognizing unit extracts the first through N-th states from an input image and calls data corresponding to the first through N-th states from the first through N-th pattern recognizing units to perform a recognizing unit.

Abstract: A method for encoding handwritten symbols operates upon penstroke data received from a device capable of sampling a stylus position at discrete intervals. Each handwritten symbol is segmented into an ordered sequence of discrete strokes. An arc length and initial and final tangent angles are evaluated for each of these strokes. Each stroke is encoded in the form of a parameter set comprising position coordinates of the initial and final endpoints of the stroke, the arc length, and the initial and final tangent angles. In specific embodiments of the invention, the segmentation is based, in part, on properties of the handwritten symbol when it is expressed as a curve .phi.(s), wherein s represents arc length and .phi. represents the net turning angle.

Abstract: A bounding box is generated for a scanned image. A vertical edge of the scanned image is detected. A horizontal edge of the scanned image is also detected. The vertical edge is used as a rightmost edge of the bounding box. The horizontal edge is used as a bottom edge of the bounding box. The vertical edge is detected by, for each scan row, detecting when a first pixel in a substep of pixels of the scan row has a grey value which differs from a grey value of a pixel immediately to the right of the first pixel by at least a first predetermined constant. When such a first pixel is found, any edge from a previous row which is continued by the first pixel is detected. A longest edge in any row continued by the first pixel a length of the longest edge is stored. Also stored is an indication of a rightmost pixel of any edge continued by the first pixel.

Abstract: A pattern inspection apparatus includes a device for acquiring image data of a target inspection pattern, and an image processing unit for extracting a defect in the target inspection pattern on the basis of the image data and image data of a reference pattern.

Abstract: The invention provides improved machine vision methods for determining a calibration relationship among the imaging reference frames of multiple cameras (or other image acquisition devices) that acquire images of a common moveable object. The calibration relationship includes parameters for transforming positional coordinates among the camera's respective imaging reference frames. The method includes placing on the common object, stage, or belt a plurality of fiducials, e.g., calibration marks, that are at known positions relative to each other with respect to a frame of reference referred to as the "alignment reference frame"; aligning an axis of the alignment reference frame with an axis of the object's reference frame of motion, i.e.

Abstract: A method of processing a matrix of gray scale picture elements derived from lifting an image of a document to extract strokes from the image of the document comprises the steps of (a) calculating a global average value of the gray scale picture elements, (b) placing a window over a portion of the matrix of gray scale picture elements, (c) calculating a local average value of the gray scale picture elements contained in the window of step (b), (d) comparing the global average value of step (a) with the local average value of step (c), (e) calculating a relative figure of importance (RFI) value for each gray scale picture element contained in the window of step (b) when the local average value calculated in step (c) is less than the global average value calculated in step (a), (f) increasing the size of the window of step (b) and replacing the window of step (b) with this increased sized window when the local average value calculated in step (c) is greater than the global average value calculated in step (a), a

Abstract: A document reading apparatus is provided with a feed roller that feeds a document in its feeding direction while making contact therewith, a rotation driving power source that rotationally drives the feed roller, and a power transmitting mechanism enabled to transmit rotational driving power from the rotation driving power source to the feed roller. Two driving paths, each having a reduction ratio different from the other corresponding to two resolutions, are provided at at least one part of the power transmitting mechanism and a planetary gear for selecting one of the driving paths as a transmitting path of the aforesaid rotational driving power is provided at a branch point of the power transmitting mechanism. Which resolution to select is determined and selection of the driving path is controlled by a selector based on the result of the determination.

Abstract: A decimation device for digital data sequences includes a multiplexer mounted in cascade with a calculator producing, alternately, during a first calculation cycle, the mean M.sub.i of at least two data sequences representing the pixels P(i,j) and P(i,j+1) in a line L.sub.i, and during a following calculation cycle, the mean M.sub.c. The mean M.sub.c is composed of the mean M.sub.i and of an intermediate sequence previously calculated and stored in a storage register arranged between an output for the calculation stage and at least one of the inputs to the multiplexer. The intermediate sequence represents the mean M.sub.i-1 of at least two data sequences representing pixels P(i-1,j) and P(i-1,j+1) in a line L.sub.i-1, where i varies from 0 to N-1 and j varies from 0 to M-1.

Abstract: The present invention relates to a device that measures and counts near surface bubbles on a copper bar, as well as the cracks on the lateral sides. The device detects said flaws in real time. The device is based on the image digitalization technique. The device consists of a personal computer, a video camera, a digitalization card, water and air sprinklers, external monitor and camera frame. The input of the images is performed with the video camera protected within the frame. The object of the water and air sprinklers is to make the bar flaws more visible. The external monitor allows to focus the area of interest and the digitalizing card within the computer, following the instructions of the program, performs the image analysis.

Abstract: In a three-dimensional reference image segmenting method and device, a two-dimensional image of a reference object and a shape data of a pattern obtained by transforming the image is stored together with depth data of the reference object in a memory as a reference pattern. On the basis of a local Fourier transform image date of an input image supplied from an image transform unit and the reference data of the reference pattern read out from the memory, a deform amount estimating unit calculates the amount of deform (displacement vector) required to make both the images coincident with each other to a possible extent. An inverse Fourier transform unit generates a deformed reference image by an local inverse Fourier transform based on the displacement vector and the data of the local Fourier transform image data of the reference pattern.

Abstract: A method to extract automatic high level features of a gray level ridge flow image including the automatic determination of the location of the high level features including the core and delta points of a fingerprint image in the form of a gray level image, comprising the following five steps:(1) determining a ridge flow direction of each gray level pixel of a fingerprint image and assign it with a direction code;(2) finding out a block directional flow and assigning each with appropriate direction code thereof;(3) correcting block flow directions in which a detection of the flow direction is affected image an input image quality in order to be immune from noises infiltrated during the acquiring process;(4) locating the delta points which are based on a corrected block direction flow diagram to locate from zero to a few delta points; and(5) locating the core points which are based on a corrected block direction flow diagram to locate one to two core points.

Abstract: A method is provided for analyzing an image comprising a plurality of image pixels to detect distortion within the image.

Abstract: The present invention is a method and apparatus for automatically detecting the characteristics of a document object placed on the platen of a digital copier or similar scanning system. The invention accomplishes the detection by processing a digitized image in a highly efficient manner using binary moments, and using the moments characterizes the shape of a boundary about the object. Once determined, the shape of the object may be employed to further process the image to automatically crop, derotate, and register the image at a predefined location in an output image suitable for rendering on a document substrate.

Abstract: A radiation image which has been recorded using an irradiation field stop and has an irradiation field is read out and an image representing the radiation image is obtained. The irradiation field is recognized and the values of the image signal components corresponding to the picture elements recognized to be outside the irradiation field out of the image signal components which make up an image signal for reproducing the radiation image as a visible image are converted according to the following formula,g(x,y)=h(t).multidot.f(x,y)+{1-h(t)}.multidot.f.sub.maxwherein f(x,y) and g(x,y) respectively represent the values of a picture element in a position (x,y) before and after the conversion, f.sub.max represents a maximum density and h(t) represents a continuous differentiable function.