Abstract: Ruled-line information is stored in a ROM. The ruled-line information relates to ruled lines present in images of originals having a plurality of different formats. Based upon consecutive pixels present in an entered image of an original, the format of this original image is identified as one format of the formats that have been stored in the ROM. Ruled lines present in the entered original image are extracted based upon the ruled-line information of the format identified.

Abstract: The method and apparatus enables any user to search for logos in document images stored in a bitmap format. The search efficiently compares bitmap or image data by extracting a series of connected components. These connected components are grouped according to region where each region may be a potential logo. Shape and density parameters of a region are determined and compared to the parameters of the stored logo image. If a region is successfully matched then that region is aligned and scaled to the corresponding stored logo image. Thereafter, a bitwise comparison is then performed between the scaled and aligned region and the logo image. A match score is assigned to each region along with other pertinent information about the region, and is stored in a ranked logo list database. The ranked logo list database represents a list of logos found in the document image.

Abstract: A bead disposition is imaged while performing illumination by reflection illumination or penetration illumination so that a peak point of brightness brighter than the outside of a bead is observed in the inside of the bead and the circumferential portion of the bead is darker than the outside of the bead. A peak point of brightness in the inside of a bead is detected from the image data and adjacent pixels are successively painted with the peak point of brightness as a starting point, wherein the adjacent pixels have a brightness level of t2 or greater. Then, the image data is binarized. A space filling percentage of beads is calculated from the number of white or black pixels of an obtained binary image. Alternatively, binarization can also be performed after detecting a circle or circular arc pattern corresponding to a dark portion in the circumferential portion of bead from the image data by a pattern matching method and painting the inside of the circle or circular arc.

Abstract: A separating position candidate detecting unit 21 detects separating position candidates in a character row derived by a character row deriving unit 11. A character candidate separating unit 31 separates character candidates by using the separating position candidates obtained by the separating position candidate detecting unit 21. A newly provided separation shape determining unit 32 determines separation shapes at the same time. A character recognition unit 41 provides a character kind and separation shape of a reference pattern most resembling each separation shape by using a character recognition dictionary 42, in which reference patterns are stored each for each separation shape.

Abstract: A bead disposition is imaged while performing illumination by reflection illumination or penetration illumination so that a peak point of brightness brighter than the outside of a bead is observed in the inside of the bead and the circumferential portion of the bead is darker than the outside of the bead. A peak point of brightness in the inside of a bead is detected from the image data and adjacent pixels are successively painted with the peak point of brightness as a starting point, wherein the adjacent pixels have a brightness level of t2 or greater. Then, the image data is binarized. A space filling percentage of beads is calculated from the number of white or black pixels of an obtained binary image. Alternatively, binarization can also be performed after detecting a circle or circular arc pattern corresponding to a dark portion in the circumferential portion of bead from the image data by a pattern matching method and painting the inside of the circle or circular arc.

Abstract: For enhanced accuracy in determining a shape in an image, a shape model can be trained for parameter settings, based on training data for which the desired shape is specified. Such training can be applied to models in which a shape is treated in a sectored fashion, with training applied to each of the sectors. The suitability of an objective function, obtained by training or otherwise, can be assessed by using near-correct shapes which are generated by randomly perturbing a ground-truth shape.

Abstract: A computer based image processing system uses an extraction process to include a pressure invariant feature for measuring distances between minutiae. The feature extraction process identifies one or more of the following features of the fingerprint: an orthogonal image contrast, a parallel image contrast, and a feature confidence. A ridge counter process, executing on the computer system, determines the number of ridges (ridge count) running across two given points and further qualifies (invalidates) this count if the confidence value of the pixels in the region adjoining the region is not reliable. The ridge count feature between minutiae is used for determining reliable features when matching fingerprints.

Abstract: A system and method for determining offset errors between line and pixel coordinates of landmarks in a digitized image generated by an imaging system disposed on a spacecraft and line and pixel coordinates predicted by a mathematical model of the imaging system using landmark geodetic coordinates on the Earth. The system and method use landmarks in symbolic form, such as perimeters of lakes and islands that are stored in a database. A digitized image generated by the satellite-based imaging system is processed to extract a patch of the image containing a landmark. The image patch is then upsampled (magnified). The landmark boundary is processed using a mathematical model of the imaging system to generate absolute coordinates of the boundary pixels, which are upsampled and rasterized to produce a landmark mask.

Abstract: In a data processing method and apparatus, an image signal representative of features in a space of two or more dimensions is progressively filtered at monotonically increasing filtering scales, the signal being input to the first filtering scale, and the input to the or each subsequent filtering scale is the output signal from the previous scale. Each input has a number of data samples each of which represents a respective region of the space, and at each filtering scale those samples are sorted into connected-sets, each of which contains a plurality of substantially equal value to data samples or a respective individual data sample not in other sets. The process then selects those connected-sets which contain data samples having values which are local maxima or local minima and which represent a composite region of a size related in a predetermined way to the current filtering scale.

Abstract: A handwritten character recognizer having an input cluster buffer and a point buffer with dynamic and static stroke feature extraction and segment analysis by conical boundaries for identification of stroke segments dynamic stroke feature extractor static stroke feature extractor. A stroke recognizer compares single copies of idealized stroke representations with hierarchically approximated multiple scaled topological representations of a current stroke, followed by stroke proportion discrimination comparing a selected topological representation of the current stroke with boundaries defined by linear combinations of features of direct and reversed ideal stroke prototypes to provide a stroke identification. A cluster recognizer maintains a time ordered current stroke buffer and previous stroke buffer and constructs a per stroke area of influence list. The time ordered buffers are scanned to generate a spatially ordered window buffer.

Abstract: A local information extracting portion 11 extracts a plurality of ridge line candidate images for each of local regions into which an image has been divided. A first ridge line candidate image selecting portion 12 selects a first ridge line candidate image that has high likelihood of a ridge line for each local region. A connectivity evaluating portion 13 evaluates the connectivity of first ridge line candidate images between each local region. A clustering portion 14 clusters local regions corresponding to the evaluated result of the connectivity evaluating portion 13 and generates a local region group. A cluster evaluating portion 15 determines an initial local region group that is a local region group having high likelihood of a ridge line. A ridge line image restoring portion 16 outputs a first ridge line candidate image for each local region that is included in the initial local region group.

Abstract: Apparatus for detecting etch defects in an object having a plurality of openings by generating an image signal of the object representing light intensity values of a plurality of pixels, processing the image signal to form data signals representing light intensity values of groups of pixels, filtering the data signals to remove signals representing the ends of the openings, and pairing two data signals when the groups of pixels represented thereby represent an etch defect.

Abstract: The present invention provides a method for reducing the variations attendant to free-form execution of a signature by providing a delimiting area in which a signature string, and/or portions of a signature string, is/are to be constrainedly transcribed. Also disclosed is a method of signature verification based on analysis of the environs attendant to the signature-string.

Abstract: The bitmap image data is analyzed by connected component extraction to identify components or connected components that represent either individual characters or letters, or regions of a nontext image. The connected components are classified as text or nontext based on geometric attributes such as the number of holes, arcs and line ends comprising each component. A nearest-neighbor analysis then identifies which text components represent lines or strings of text and each line or string is further analyzed to determine its vertical or horizontal orientation. Thereafter, separate vertical and horizontal font height filters are used to identify those text strings that are the most likely candidates. For the most likely title candidates a bounding box is defined which can be associated with or overlaid upon the original bitmap data to select the title region for further processing or display. Captions and photographs can also be located.

Abstract: Methods and apparatus for removing scratch and wire noise from digitized images handle the difficult case of non-uniformly shaded images which is the common situation in real applications. The methods and apparatus also utilize a soft edge noise mask to allow a smoother blending of the reconstructed noise and the image around the noise. wire noise removal from digitized images wherein a scratched portion of a digitized image which requires repair or noise removal is identified and a binary mask is generated which distinguishes the defined scratched portion from the other portions of the digitized image. The methods and apparatus utilize both the spatial and frequency domains to effectively reconstruct many contiguous noisy pixels, reconstruct textures even when they are large featured, maintain sharpness and maintain continuity of features across a noisy region. A set of instructions for executing the methods is contained in a program storage device.

Abstract: Apparatus for detecting activation signals in functional MRI images comprises MRI apparatus for deriving MRI volume data temporal signals for voxels in image slices, the data being grouped in segments respectively corresponding to active and inactive periods.

Abstract: An image forming apparatus includes a reading unit for reading an image on a document, a storage unit for storing image data output from the reading unit, an image forming unit for forming an image on a recording medium on the basis of the image data stored in the storage unit, a discrimination unit for discriminating whether or not a hole is present in the document read by the reading unit by analyzing the image data stored in the storage unit, and a control unit for, when the discrimination unit determines that the hole is present, controlling the image forming unit to form an image while masking an image in an area where the hole is present.

Abstract: A method and apparatus for comparing symbols extracted from binary images of text for classifying into equivalence classes. The present invention uses a Hausdorff-like method for comparing symbols for similarity. When a symbol contained in a bitmap A is compared to a symbol contained in a bitmap B, it is determined whether or not the symbol in bitmap B fits within a tolerance into a dilated representation of the symbol in bitmap A with no excessive density of errors and whether the symbol in bitmap A fits within a tolerance into a dilated representation of the symbol in bitmap B with no excessive density of errors. If both tests are passed, an error density check is performed to determine a match. The dilated representation of the bitmap accounts for various quantization errors that may occur along the boundaries of a symbol defined in the respective bitmaps. The dilation utilized preserves the topology of the symbol.

Abstract: Pattern recognition, particularly character recognition, is effected by categorizing a cipher by a unique description using three feature sets identifying the type and location of lines which form the cipher, the location of the endpoints of each of the lines, and the location of the intersections of the lines. In the process according to the invention, lines are located by expanding or contracting the original image of the cipher to fit into a sparse space pixel matrix. The image is then thinned to define lines and so as not to obliterate lines which give the cipher its uniqueness. Once the cipher is thinned, fuzzy logic is used to "locate", that is, define a set of coordinates for, each line relative to an x-y coordinate system. (As each line is located, it is removed from the matrix image to facilitate locating the remaining lines.) After all the lines are located, the endpoints of each line are located and the intersect points are located.

Abstract: A technique for object information extraction from images which retains fuzziness as realistically as possible. The technique is used for image segmentation of fuzzy objects for n-dimensional digital spaces based on the notion of "hanging togetherness" of image elements specified by their fuzzy connectedness. A specified fuzzy object is extracted and all fuzzy objects present in the image data are identified by segmenting the image using the notion of "hanging togetherness" of image elements specified by their fuzzy connectedness as defined herein. The technique is used in a preferred embodiment to quantify MS lesions of the brain via magnetic resonance imaging.

Abstract: A process and apparatus for the production of line data from measurements correlated with specific characteristics of an object measured. The invention provides that the measurements thus obtained are coded by measurement points and stored in an orientated matrix, the individual elements of this matrix are examined repeatedly for prescribed correlations and the elements meeting in each case the prescribed correlation are in each case coordinated into groups of point coordinates defining a line plot, and the line plots thus produced are depicted.

Abstract: Image processing apparatus is disclosed for splitting, for subsequent storage or processing by OCR apparatus, character images in digital form comprising connected characters, the image processing apparatus extracts an image skeleton from an input image and represents the topology of the image skeleton in a first data structure; generates, from the first data structure, second data structures representing the topologies of the results of possible splits of the input image skeleton; evaluates the second data structures and selects one of the possible splits to thereby split the input image.

Abstract: A method and apparatus for recognizing a script written character wherein the character is entered using character entering means and digitised by appropriate means. The digitised character is then stored in, for example, a memory. Codes representing topological and vector features of the character are extracted from the character, then the topological and vector features of the character are compared with topological and vector features of a plurality of reference characters defining a set of reference characters stored in a memory. Each of the reference characters included in the set corresponds to a specific script written character. A logic process is then performed to determine which reference character of the set of reference characters has topological and vector features most closely corresponding to the topological and vector features of the digitized character thereby identifying the script written character.

Abstract: From an image memory of a fingerprint picture, gradient vectors of the picture are calculated. And from distribution of gradient vectors in a subregion, ridge direction of the subregion are determined, and confidence of the determined direction is also defined.

Abstract: This invention relates to a method and apparatus for recognizing a script written character. The character is entered using character entering device and digitised by appropriate device. The digitised character is then stored in, for example, a memory. Codes representing topological, vector dimension features and the microfeatures of the character are extracted from the character, then the features of the character are compared with a set of reference features corresponding thereto stored in a memory. Each of the set of reference characters corresponding with a specific script written character. A logic process is then performed to determine which of the set of reference features most closely corresponds to the topological features of the digitized character thereby identifying the script written character. The relative weighting of the feature can be varied for different types of script or confusing characters to enable still more accurate recognition.

Abstract: A three-dimensional model is formed from information of two-dimensional drawings. Information representing section is formed from the two-dimensional drawing information. Geometrical and topological information is then extracted, and the three-dimensional model is thereby formed.

Abstract: When a leading end of a run on an upper row in a binary mask which performs raster scan over a binary image is detected, a run label which was assigned to the same run when scanning a row preceding the current row by one row is read out, and label diversion destination data is read from a concatenation table using the read run label as an address. A temporary label is determined by comparing a concatenated label, which indicates the minimum label value of the runs prior to the run on the upper row of which leading end is detected and adjacent to the run existing on the run on the lower row of the binary mask, with the diversion destination data read from the concatenation table. If one of them is 0, the other label data is issued as the temporary label. If both of them are 0, a new label is issued as the temporary label. If the compared label values are different from each and are not 0, the smaller label value is written into the concatenation table using the larger label value as the address.

Abstract: Two-dimensional image data of an object is processed to inspect a circular arc configuration of the object. Initially, the two-dimensional image data is subjected to a data extraction process in which image edge points of the circular arc configuration of the object are extracted from the two-dimensional image data. The image edge points are then applied to a process to approximate the circular arc configuration of the object by defining a circle which minimizes a mean value of a fourth power error of the image edge points relative to the circle, the thus defined circle being used to approximate the circular arc configuration of the object.

Abstract: This invention relates to a method and apparatus for recognizing a script written character. The character is entered using character entering means and digitised by appropriate means. The digitised character is then stored in, for example, a memory. Codes representing topological features of the character are extracted from the character, then the topological features of the character are compared with a set of reference topological features stored in a memory. Each of the set of reference characters corresponding with a specific script written character. A logic process is then performed to determine which of the set of reference features most closely corresponds to the topological features of the digitized character thereby identifying the script written character.

Abstract: Pattern recognition, for instance optical character recognition, is achieved by defining a minimal bounding rectangle around a pattern, dividing the pattern into a grid of boxes, comparing a vector derived from this partitioned pattern to vectors similarly derived from known patterns, choosing a set of Pareto non-inferior candidate patterns, and selecting a recognized pattern from the set of candidates. The vectors include pixel density matrices, matrices of horizontal connectivity of boxes, and matrices of vertical connectivity of boxes.

Abstract: A device for coding still images has a device for segmenting the images into homogeneous regions whereto a specific label is attached, and a device for coding the contents of the contour. The contour coding device has a sub-assembly which detects contours for each regions, for each region successively considered, it detects points which are to form a series of control points and transforming the part of the contour joining two successive control points into a straight segment. Then the validity criterion is verified. If the criterion is not met, at least one supplementary control point is introduced between the two control points and so on until it is. The transformation is applied until the contour loop is completely closed.

Abstract: An image recognition device includes an input section for inputting an image; a pattern extraction section for extracting a plurality of predetermined patterns from the input image; a first feature extraction section for extracting a first feature indicating the direction in which the first pattern of the extracted predetermined patterns exists with respect to a point on a second pattern of the extracted predetermined patterns; a second feature extraction section for extracting a second feature indicating an amount of the first pattern existing in the direction with respect to the second pattern, based on the first feature; and a recognition section for recognizing the input image based on the second feature.

Abstract: An optical character recognition system which automatically reads handwritten characters and the like which do not have to be printed in a special format. Recognition tables derived from the pattern bit map and from a skeleton pixel matrix describe the character in terms of the relative position of the pixels along the borders of the character and provide a plurality of recognition strings, one in each table, for the front and rear views of the character and for the shape of the holes in the character which are opened from the top (as in the numeral four) or opened from the bottom (as in the numeral seven). From the recognition tables, the characters are recognized by searching recognition files containing blocks of successions of lines of code corresponding selectively to the codes in the recognition tables.

Abstract: A recognition system for identifying members of an audience, the system including an imaging system which generates an image of the audience; a selector module for selecting a portion of the generated image; a detection means which analyzes the selected image portion to determine whether an image of a person is present; and a recognition module responsive to the detection means for determining whether a detected image of a person identified by the detection means resembles one of a reference set of images of individuals.