Abstract: A method and apparatus locates edges of patterns in a stored image, such as edges of bars or finder patterns of machine-readable symbols. The present invention first selects an edge point or pixel within a stored image. A dominant direction is determined therefrom by examining a region surrounding the selected point. A smaller window is then examined around the edge point to derive a hexadecimal code based on binary values of pixels which surround the edge point. The hexadecimal code is compared to a table of codes, where each code maps to a similar window and identifies a next edge point. The process continues in the dominant direction until the entire edge is located. Circuitry for rapidly performing the its location of patterns in a stored image, such as ages of bars of finder patterns of machine-readable symbols.

Abstract: An arrangement is described for determining the skew angle of a document image without requiring text or other skew detection information in the document or without human intervention. The arrangement accomplishes this by determining an edge of the document image within a scan image and uses that edge to determine the skew angle of the document image. The edge can be determined by locating the first or last document image pixel of each scan line of pixels in the scan image that belongs to the document image. This is accomplished by comparing each pixel of each scan line of pixels with a predetermined scan line of background pixels. The skew angle of the document image is then determined by computing the slope of the detected edge in the scan image. An arrangement of determining the boundary of the document image is also described.

Abstract: An electronic still camera is provided with an electronic image sensor for generating an image signal corresponding to a still image of a subject and an orientation determination section for sensing the orientation of the camera relative to the subject. The orientation determination section provides an orientation signal recognizing either the vertical or the horizontal orientation of the camera relative to the subject. An image processor is responsive to the orientation signal for processing the image signal and correcting the orientation thereof so that the still image is output from the image processor in a predetermined orientation. In this way, the electronic still camera can be positioned in a variety of orientations relative to a subject, including both clockwise and counterclockwise vertical "portrait" orientations and a horizontal "landscape" orientation, without affecting the orientation of the images output by the camera.

Abstract: There is provided a stamp-making method and apparatus therefor for making a square or circular stamp which can be accurately imprinted with reference to a top-indicating mark provided thereon, without discriminating the orientation of the stamp image engraved on the stamp surface thereof. There is formed first direction-oriented image data having each of a plurality of component elements of plate-making image data oriented in a first direction. When it is detected that it is difficult to discriminate orientation of the stamp body having the mark without reference to the mark, an instruction is given for arranging the image data in a manner oriented in a second direction. According to the instruction, the first direction-oriented image data is converted to a second direction-oriented image data in which the component elements of the plating-making image data are oriented in the second direction. A stamp body is engraved by the use of the second direction-oriented image data to thereby make a stamp.

Abstract: An image processing apparatus is capable of rotating a read image into a desired direction according to the content of a mark sheet to be read prior to the originals. As a result, the operator need not pay particular attention to the direction in which he sets the originals on the reader unit. To this end, a detection unit detects an identifier on, or attached to, the original on the reader unit, a rotation unit rotates the read image, and an output unit for releasing the rotated image in accordance with the content of the mark sheet, for example by rotating the image in accordance with the receiving equipment.

Abstract: The present invention is a method and apparatus for the fast rotation of binary images represented as a run length encoded data stream. The system employs a cache memory for efficiently representing parsed RLE data in a rotated orientation, where only the white-black transition points are recorded. Subsequently, the cache memory is further processed using a logic operation to fill in the runs between the recorded transition points during concurrent transfer of the image data to a band of page buffer memory. The process is repeated for a plurality of bands to complete the rotation/decoding process for an entire bitmap image.

Abstract: An image processing device includes a reader for reading an original image and generating image data; a display for displaying a plurality of screens for selecting a desired image edit mode from a plurality of image edit modes, each of the screens having image edit modes which are different between each screens; a display control for selecting a screen among the plurality of screens in response to the type of original image and causing the display to display the selected screen; and image editer for performing image processing on the image data outputted in response to an image edit mode selected through the selected screen.

Abstract: An image processing system which allows a variety of different types of split-screen image data, including window displays, to be generated. The image processing system of the present invention further can display windows of various shapes to ensure diversified use of display screens. The image processing system also enables various types of movement and rotation transformation as well as scaling transformation to be performed while reducing the calculation load on the CPU.

Abstract: The present invention is a method and apparatus for the processing of images that have been compressed using a discrete cosine transform operation, and particularly JPEG compressed images. In a preferred embodiment, the rotation of image blocks is accomplished by sign inversion and transposition operations to accomplish intrablock operations. Subsequently, one of a number of alternative methods is employed to accomplish the same image processing on an interblock level, thereby enabling the rotation or mirroring of compressed images. The two stage process allows the use of either a standardized JPEG system with enhancements or a hybrid processing method, thereby accomplishing the image processing in conjunction with compression or decompression operations and minimizing the need for large memory buffers to accomplish the image processing.

Abstract: A copying machine includes an original feeding unit for feeding an original to a platen, a reading unit for reading an image of the original fed by the original feeding unit, a reducing unit for reducing the image supplied from the reading unit, a storing unit for storing a plurality of images supplied from the reducing unit in a predetermined area, an input unit for entering information regarding an orientation of the images of the original, a control unit for use in changing the area where the image is to be stored by the storing unit depending on the information regarding the orientation supplied from the input unit, and a recording unit for recording a plurality of images stored in the storing unit on a single sheet.

Abstract: In a telemanipulation system for manipulating objects located in a workspace at a remote worksite by an operator from an operator's station, such as in a remote surgical system, the remote worksite having a manipulator with an end effector for manipulating an object at the workspace, such as a body cavity, a controller including a hand control at the control operator's station for remote control of the manipulator, an image capture device, such as a camera, and image output device for reproducing a viewable real-time image, the improvement wherein a position sensor associated with the image capture device senses position relative to the end effector and a processor transforms the viewable real-time image into a perspective image with correlated manipulation of the end effector by the hand controller such that the operator can manipulate the end effector and the manipulator as if viewing the workspace in true presence.

Abstract: A method and an apparatus for rotating images on a computer system is disclosed. The system includes a processor for accessing an image frame buffer using a set of CPU addresses, a memory for storing the frame buffer, a controller for directing the pixel data of the frame buffer to a display device using a set of controller addresses. The two sets of addresses are not necessarily the same. In fact, numerous advantages could be had from manipulating those two sets of addresses resulting in image rotation operations for the display device.

Abstract: In a character recognition system, a method and apparatus for correcting the skew of a document image. Skew correction is typically performed during segmentation of the document image into text and non-text parts. Skew correction generally involves skew angle determination and correction of the document image based on the skew angle. A skew angle is determined through the steps of: providing a set of associated rectangles representing the document image, identifying a column edge associated with the set of associated rectangles, comparing rectangles from the set of associated rectangles to identify those that are in the same column and suitably far apart, calculating a tangential angle between the rectangles identified and identifying the most common tangential angle as the skew angle.

Abstract: A method for Golden Template Comparison (GTC) is provided that can be used to efficiently perform flaw and defect detection on a two-dimensional test image that is at least rotated and/or scaled and/or sub-pixel translated. Run-time inspection speed and accuracy is substantially improved by retreiving a golden template image that is rotated and/or scaled and/or translated in a manner substantially similar to the test image. This is accomplished by storing, in an array, a varied plurality of golden template images, each golden template image being characterized by a different combination of at least rotation and/or scale and/or sub-pixel translation. The array is indexed by the respective quantized rotation and/or quantized scale and/or sub-pixel translation of each version of the golden template image. The array can be either one-dimensional or multi-dimensional.

Abstract: Registration of images includes rotational and translation correction. Rotational correction includes estimating a rotational-correction angle by polar K-space cross-correlation. K-space regridding is used to rotate an image by the estimated correction angle. Translational correction is obtained by K-space phase-plane conjugation.

Abstract: The image processing system according to this invention is very practical, because it allows a variety types of split-screen image data, including window displays, to be generated. In addition, the image processing system according to this invention, which can display windows of various shapes, ensures diversified use of display screens. This invention also enables various types of movement and rotation transformation and scaling transformation while reducing the calculation load on the CPU.

Abstract: An image synthesizing device in which partial images can be rapidly processed (for example, by enlarging, reducing, turning) and synthesized. When synthesizing information for partial images stored in a partial image storage portion is calculated by a synthesizing data calculating portion, accessible rectangular areas, which consist each of a plurality of pixel rows and columns, of partial images stored in the partial image storage portion are accessed one by one by a partial image accessing portion and each of accessed areas is processed by enlarging or reducing or turning on the basis of the calculation results given from the synthesizing data calculating portion. The processed partial images whose areas have been enlarged, reduced or turned, are synthesized by a partial image synthesizing portion on the basis of the synthesizing information, and a synthesized image is stored in a synthesized image storage portion.

Abstract: A method for detecting an amount of rotation or magnification in a modified image, includes the steps of: embedding a marker image having a pair of identical features separated by a distance d and oriented at an angle .alpha. in an original image to produce a marked image, the marked image having been rotated and/or magnified to produce the modified image; performing an autocorrelation on the modified image to produce a pair of autocorrelation peaks corresponding to the location of the features of the marker image in the modified image; and comparing the separation d' and orientation .alpha.' of the autocorrelation peaks with the separation d and orientation .alpha. of the features in the marker image to determine the amount of rotation and magnification in the modified image.

Abstract: An Electron Beam Tester which corrects deformation of a secondary electron image produced from scanning a sample with an electron beam. The secondary electron image is stored in a storage unit. Luminance data of the stored image is accumulated to obtain a projected luminance distribution. The projected luminance distribution data is then analyzed by a parallelism evaluation unit to obtain a rotation angle. Then the rotation angle is used to determine maximum parallelism and correct deformation of the secondary electron image by providing deflection control which transforms the deflectors .

Abstract: A personal imaging computer system, which is connectable to and operable with a computerized local or wide area network, identifies characters in a document on which the characters are formed. The system scans the document to obtain a gray-scale image of the document, de-skews the gray-scale image, generates a binary image from the de-skewed gray-scale image by comparing the gray-scale image with the threshold, segments the binary image to locate individual characters within the binary image and to determine the shape of the individual characters, extracts gray-scale image information from the gray-scale image for each such individual character based on the location and shape of the character in the binary image, recognition-processes the extracted gray scale image information to determine the identity of the character, and stores the identity of the character.

Abstract: A binary image coordinate transformation method can simplify the transformation processing for those portions consisting of pixels having the same single color value, thereby increasing the processing speed of binary image transformation overall. The original image is divided into a plurality of sub-regions. For each sub-region, it is detected whether or not the sub-region consists of pixels having the same color value of two color values of a binary image. The post-transformed image is divided into a plurality of sub-blocks. For each sub-block, all the sub-regions containing pixels corresponding to those pixels included in a sub-block are detected. When it is determined that the all the detected sub-regions corresponding to the sub-block have the same color value, the sub-block of the post-transformed image is covered with that single color in one go.

Abstract: The present invention relates in general to optical scanning and image processing, and relates more particularly to a document imaging system which detects skew and/or detects size/shape of a document image. Preferred embodiments utilize a background with optical characteristics which contrast with those of the scanned document. In one embodiment, a document imaging system generates scanning signals in response to optical characteristics of a medium such as a sheet of paper, detects transitions in the scanning signal which define points along one or more edges of the medium, establishes a skew angle between the detected edges and a reference orientation, and compensates for skew by modifying the scanning signals as required such that the resultant skew angle is substantially equal to zero.

Abstract: A first station encodes data as data symbols, rotates the data symbols so that data symbols corresponding to one another across sync portions of data segments of a frame are arranged to be processed together, and inserts the rotated data symbols into data portions of the data segments of the frame such that the frame has a plurality of data segments, wherein each data segment has a segment sync portion and a data portion. A second station processes the data symbols so that corresponding data symbols in a first set of data symbols are processed with a twelve data symbol delay and corresponding data symbols in a second set of data symbols are processed with a twenty-four data symbol delay depending upon the position of the data symbols in relation to the sync portions.

Abstract: This invention provides a device and method to detect a degree of misregistration and angle of misrotation of an object image, a device to measure an object using such a degree of misregistration and an angle of misrotation and a device to record data representing a model. A model of the object used is imaged and at least two regions are established in the image data of the model. Delimiting values which are extracted from geometric features of these regions are stored as reference values. The same regions are then extracted from an image of the object, and similar delimiting values are obtained from their geometric features. The conditions so obtained are compared with the stored standard reference values. The degree of misregistration of the object to be processed with respect to the object used as the standard of comparison is calculated from the difference between these two values.

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: The present invention is a method and apparatus for the rotation of images in conjunction with a block-wise, variable-length data compression operation. In a preferred embodiment, the rotation of image blocks on a microscopic level is accomplished independently from the rotation of the blocks themselves (macroscopic), stored in electronic precollation memory and the rotated image produced upon decompression of the stored rotated data.

Abstract: A blank space is provided at an appropriate position on a recording paper as determined by either input image data to be recorded, or the size of the recording paper.

Abstract: This invention uses an optical imaging system input to a computer to synthesize regularly spaced reference and test line patterns replicated from binarized edges taken from the optical images of structures to be aligned. The two patterns are then combined resulting in an interference pattern in which the spacing of moire fringes indicates the amount of movement of the structures required to bring them to a predetermined alignment. All of the procedures are observable on the interference display. These moire fringe pattern are a quantitative measure of angular alignment between the reference and test images and of the edges from the structures from which they were taken. An operator programmed and controlled template automatically processes moire edge alignment measurements.An automatic template processor generates a test sequence template from the chosen areas for measurement by the operator.

Abstract: An image processing apparatus comprises an input unit for inputting image data, an extracting unit for extracting outline data on the basis of the inputted image data, a deriving unit for performing size change, rotation, and movement processes to the extracted outline data and for deriving outline data corresponding to an image to be outputted, a forming unit for forming image data on the basis of the derived outline data, and an output unit for outputting the formed image.

Abstract: An optical image reconstructing apparatus includes an optical image forming system. A focused surface controller moves the position of an object surface focused by the optical image forming system. An image pickup converts an image of an object formed by the optical image forming system into an electrical signal and outputs the electrical signal as an image signal. An adder adds a plurality of image signals output from the image pickup while the position of the object surface is moved by the focused surface controller. A frequency filtering device performs a frequency filtering operation on an image signal outputted from the adder to produce a filtered image signal. A rotation controller rotates the object or the optical image forming system at a predetermined angular interval around one point on an optical axis of the optical image forming system.

Abstract: A method and apparatus for use in automatic determination of the denominations of currency bills employs a light scanner to produce pixel signals representative of light pixels from bill surface portions extending across bill sides The pixel signals representative of side and top edges of the printing on the bill surface are found and used to generate a deskewed array of pixels representative of bill corner images that include the bill's denomination numbers Pixel signals indicative of light passed through bill portions enable the detection of a security thread in a bill A hierarchial technique for determining bill denominations is described

Abstract: A series of documents to be processed are optically inspected, and compared with a reference standard to identify the relative orientation of each document. To this end, an image acquired from the document (preferably from each side of the document) is focused on an array of picture elements (pixels) for electronically converting the acquired image to digital form. This produces a digitally encoded image, which can be enhanced if desired, defined by an array of pixels corresponding to the acquired image. For each acquired image, two reference areas, which are preferably symmetrically located on the document, are inspected for the presence of a pre-selected reference mark. Location of the reference mark in one of the two reference areas operates to determine the orientation of the document.

Abstract: A method for rotating bi-level image data plus or minus 90.degree.. An 8.times.8 pixel block is selected for rotation from the image data and, using lookup tables, an intermediate matrix is generated dynamically and stored into two 32-bit registers in the processor. Generation of the intermediate matrix partially rotates and stores the odd-numbered rows from the block of data in the first 32-bit register and the even-numbered rows from the block in the second 32-bit register. The first and second registers are copied, respectively, into third and fourth 32-bit registers. The first and second registers are then shifted one bit in opposite directions. The first register is logical ORed with the fourth register, and the second register is logical ORed with the third register so that the even bits are replaced in one register and the odd bits in the other. These steps are repeated for all remaining blocks of image data.

Abstract: A technique for compensating for egomotion of the camera used to record a pair of two-dimensional views of a scene when the pair of images is to be used to provide a three dimensional representation of the scene. The technique involves comparing histograms of the intensity levels of pixels of corresponding epipolar lines in the pair of images for assumed amounts of egomotion to identify the amount that results in the smallest total of the sums of squared differences of the histograms.

Abstract: In a telemanipulation system for manipulating objects located in a workspace at a remote worksite by an operator from an operator's station, such as in a remote surgical system, the remote worksite having a manipulator with an end effector for manipulating an object at the workspace, such as a body cavity, a controller including a hand control at the control operator's station for remote control of the manipulator, an image capture device, such as a camera, and image output device for reproducing a viewable real-time image, the improvement wherein a position sensor associated with the image capture device senses position relative to the end effector and a processor transforms the viewable real-time image into a perspective image with correlated manipulation of the end effector by the hand controller such that the operator can manipulate the end effector and the manipulator as if viewing the workspace in true presence.

Abstract: A computer-implemented method of transmitting images from a transmitter to a receiver (e.g. in a teleconferencing application). A receiver maintains an image in a local storage (e.g. that from a previous frame in a sequence of frames) and the transmitter receives an updated image for a next temporal period (e.g. the next frame). The transmitter divides the updated image into blocks and comparing a rotating pixel sample(s) of each of the blocks from the updated image with a sampled pixel from a local copy of a receiver's image at a same spatial position of the pixel sample(s). The transmitter determines a difference between the rotating sampled pixel of each of the blocks from the updated image and the local copy of the receiver's image. It stores a reference to the block and associates the difference with the reference.

Abstract: A system and method is provided for aligning and analyzing multiple images using Glass patterns. A Glass pattern is a Moire-like pattern obtained by superimposing an image on a transformed version of itself. The method includes superimposing the images to produce a first Glass pattern which has a determinable reference point, repositioning at least one of the images to produce a subsequent Glass pattern having a different reference point, and simultaneously aligning the reference points of each of the Glass patterns so as to align the images. The reference points are determined by any of a number of methods including the use of Fourier Transforms. In various embodiments of the invention, information obtained from the nature and quality of the Glass patterns produced is utilized for image analysis without aligning the images.

Abstract: A circuit for performing high-speed 90 and 270 degree rotations of an uncompressed image in a page buffer. The rotation is performed as the data is written into and read from the page buffer, thus requiring no additional time. A special memory mapping technique is used to make efficient use of the page buffer memory.

Abstract: A technology for outputting an image, which has been subjected to a diagonal process, in such a manner that said image can be included within a recording paper sheet. Specifically, an inputted image of an original document is subjected to an enlargement (or a contraction) process and a diagonal process, followed by changing the read start address to be read from a shift memory in accordance with the size of the output recording paper sheet and that of a diagonal image to output a portion of the diagonal image, that protrudes the output recording paper sheet, into a marginal portion of the output recording paper sheet.

Abstract: Disclosed is a fast and high resolution technique for estimating the skew angle in text document images. Detection of the skew angle is an important step in text processing tasks such as optical character reading (OCR) and computerized filing. The proposed method is based on the application of a recently developed algorithm for estimation of angle of straight lines in a scanned image. The Subspace-Based Line Detection (SLIDE) algorithm formulates the multi-line fitting problem in a special parameter estimation framework such that a signal structure similar to the one in the field of high-resolution sensor array processing is obtained. Then the well-studied techniques in that formalism (e.g., the ESPRIT algorithm) are exploited to produce closed-form and high resolution estimates for line parameters.

Abstract: A three-dimensional object movement and transformation processing apparatus has a shape storing section for storing shapes of a plurality of three-dimensional objects, an affine matrix calculating section for calculating transformation affine matrices Mow indicating movement or transformation of the objects from shape prescription coordinate systems prescribing the shapes of the objects to a world coordinate system, an inverse matrix calculating section for calculating inverse matrices Mwo of the matrices Mow indicating movement or transformation of the objects from the world coordinate system to the shape prescription coordinate systems, and an inter-object processing section for changing shapes of a pair of objects overlapping with each other to processed shapes according to the matrices Mwo and Mow of the second objects to reflect an overlapping condition in the processed shapes.

Abstract: Method and apparatus for providing interpolated image data from original image data represented in terms of pixels, each pixel defined in terms of value and position in an original image, includes an original image input receiving original image data from an original image source; a page memory operatively connected to the input for storing a page of original image received; a source of interpolation parameters indicating: a slow scan initial pixel value Xinit, a fast scan initial pixel value Yinit, a fast scan x offset value FSx, a fast scan y offset value FSy, a slow scan x offset value SSx, and a slow scan y offset value SSy; a bilinear sequencer calculating for each new pixel, from the received parameters a reference pixel within the image, and a pair of interpolation coefficients for interpolating new pixel values; a memory controller retrieving to an interpolation calculator from the page memory a set of original image pixels including the pixel at the reference position, and three other pixels whose po

Abstract: A method of reversing all of the pixels in an image word uses recursive steps of shifting the image word a varying, predetermined number of spaces, followed by filtering the shifted and unshifted words through a varying, predetermined mask. The number of cycles necessary to reverse an image word having 2.sup.n pixels is n cycles, plus one additional shifting step. This method of reversing a single image word is used to generate the mirror image of an input image. This image word reversal method is also useful for rotating an image word 180 degrees about an axis through its center.

Abstract: A method of determining unintended inclination in an image of characters that is to be processed by a device such as a copy machine, a facsimile machine, or an optical character recognition system. An image of characters is generated, and expansion is performed on this image the optimal number of times. The number of expansions is based on the differences in the area of the non-background portion of the image at various stages of expansion. Next, the major direction of inclination is determined for the image. The inclination angle of the image is then determined based on the inclination factors of the character lines for various zones in the image.

Abstract: A handwritten test symbol, such as an alphabetic character, a word or a signature, written on a digitizing tablet is normalized by comparison with a model symbol to determine the transformation necessary to best fit the test symbol to the model symbol. Such transformation is then applied to normalize the test symbol. Shape information in the test symbol is preserved during such normalization. In accordance with one aspect of the invention, the model symbol is a line segment. In accordance with another aspect of the invention, the model symbol is an example of the symbol being normalized. Such normalization can be used as a preprocessing step in applications such as character recognition, text recognition or signature verification.

Abstract: A method and apparatus for making autostereographic images wherein a computer is provided with a number of planar images in digital form, in a two-dimensional array of columns and rows of digital data. The columns are interleaved and rotated to form an interleaved and rotated image so that the interleaved columns of the rotated image are substantially unaligned with the rows and columns of the digital representation of this image. The digital representation of the image is rotated using an algorithm that is independent of resolution and number of input images. The resulting image is then output on a high resolution imaging device so that the columns of interleaving are substantially unaligned with both the direction of printing and the normal to direction of printing. The output from the high resolution imaging device is then placed in a registered relationship with a means for viewing selected ones of the input images, such as a lenticular or barrier screen.

Abstract: A method and computer implementation are disclosed for correcting skew and shift which may occur as the image on a document or other field is scanned with line scanning apparatus. By selecting scans, lines are logically overlaid upon a memory map of the scanned image and a best-fit line is calculated through left-most image points near the left ends of the lines. The best-fit line substantially defines the left margin. The angle which complements the angle between the overlay lines and the best-fit line is calculated and the image map is rotated in storage, aligning the image with scanner's x-axis and y-axis. The method may be started while scanning is still in progress and the last portion of the image may be rotated according to the invention as it is being stored. The left margin is more accurately found by expanding the search for left-most pixels using search areas around the fit originally found left-most points.

Abstract: A method and apparatus for adjusting a marker in preparation for cutting that is adapted for use with fabrics having a stripe or plaid patterns that provide for an axis of symmetry. The present invention is characterized by computer assisted design matching that provides for either manual or automatic alignment of the garment segments in the computerized marker with reference and match locations on the fabric web.

Abstract: A system for determining the fine angular orientation of a bar code symbol in a two-dimensional pixel image. A window is selected from the two-dimensional image for processing. Edge magnitude information and edge direction information are determined for a plurality of pixels in the selected window. For each pixel in the plurality of pixels, the edge magnitude and the edge direction of each pixel is associated with at least one of a plurality of directional ranges if the edge magnitude associated with the pixel exceeds a threshold. For each directional range in the plurality of directional ranges, a density value associated with each directional range is determined in accordance with the number of pixels associated with the directional range, an angular mean value is determined in accordance with the edge directions associated with the directional range, and a noise value is determined in accordance with the edge directions associated with the directional range.

Abstract: In a character recognition system, a method and apparatus for segmenting a document image into areas containing text and non-text. Document segmentation in the present invention is comprised generally of the steps of: providing a bit-mapped representation of the document image, extracting run lengths for each scanline from the bit-mapped representation of the document image; constructing rectangles from the run lengths; initially classifying each of the rectangles as either text or non-text; correcting for the skew in the rectangles; merging associated text into one or more text blocks; and logically ordering the text blocks.