Abstract: A method for resolution conversion for re-sampling anti-aliased images is disclosed which decreases bandwidth costs associated with anti-aliased line art and other costs associated with interpolating these images to a desired resolution. The present method first involving the receipt of an image which is comprised of bitmap data including at least a plurality of gray-scale pixel tiles that define the image. Then receiving the image data at a first resolution and extracting pixel tile information of the received image at a second resolution. The method has the step of next using loose gray scale template matching on each of the pixel tile information with at least one of a plurality of templates so as to generate pixel-wise looseness interval values there between. Then, outputting a portion of the enhanced pixel tile information wherein the enhanced pixel tile information is formed by a matching of a template with pixel-wise looseness values.

Abstract: An increasing nonlinear-filter represented as a plurality of basis elements along with the filter values at those basis elements by finding, for each output value k of the filter, all possible observations X that result in an output value k or greater. The set of all values that map to an output value k or greater have at least one minimal element at these are known in the art as basis elements. These basis elements are arranged in a list, numbered 1 through M. Constructing a table in which basis element number corresponds to the filter output value at that basis element. Thereafter, each of the N samples (X1, . . . , XN) is inspected in turn. For each sample X1, testing each of its possible values with its respective interval components wherein, if the M basis elements are A1, . . . , AM and if basis element number j is being inspected, then each value of X1 is tested against Aij. For each value of X1=t1, . . .

Abstract: An adaptive image segmentation system and methodology based on Mixed Raster Content (MRC) format. A L*a*b* color image is processed into an object-based MRC representation. By using L*a*b* data, an expectation-maximization algorithm is used to estimate a mixture of two 3-D Gaussians, with one Gaussian representing the background pixels and the other the foreground pixels. A resultant-quadratic decision surface is calculated and all image pixels are compared against it. Depending on which side of the decision surface any given pixel falls, that pixel goes to either the background or foreground plane. The pixel-by-pixel decisions are used to comprise a mask plane. The mask plane is converted into run lengths, which are “cleaned”, and regions are merged. Large connected components are reserved as windows and are used to mask out portions of the foreground. The result is a background plane, a mask plane, a foreground plane and any number of foreground/mask pairs, consistent with the ITU T.

Abstract: A Document Image Segmentation method is disclosed for matching a plurality of templates with a received image wherein the received image being bitmap data including at least a plurality of gray-scale pixel tiles that define the received image, the matching method having the steps of first receiving said image data so as to extract pixel tile information of said received image wherein said pixel tile information being of a predetermined matrix size; then matching loosely said pixel tile information with at least one of a plurality of templates so as to generate pixel-wise looseness interval values there between; and finally outputting an identifier associated with the matching template such that said identifier indicates a classification. The classification is preferably based on at least one of continuous tone pictorials, text, half tones, high/low frequency range; and line art graphic. Also, preferably the classification contains information about the image source.

Abstract: A method is presented for encoding auxiliary information in digital images which enables the encoding of both compact auxiliary information specific to a given image and a secret message into a given image such that the hidden message contained therein is only revealed in such a way that neither the image itself nor the auxiliary information alone are sufficient to reveal the secret message. Specifically, a method for encoding an auxiliary message in a digital image, is presented having the first step of creating a set of Training Vectors from all N×N windowed observations of a digital image (D) and corresponding pixels of a message (I). Then, statistically estimating a function T from the set of Training Vectors which map the N×N windowed observations of the digital image (D) to the corresponding pixels of the message (I); and writing out the statistically estimated function T as a representation of the hidden message.

Abstract: What is presented is a method for rendering gray half-toned images from a received image wherein the received image is bitmap data including at least a plurality of gray-scale pixel tiles that define the image. The method includes the steps of first receiving a gray image so as to extract pixel tile information of the received image. The next step is matching loosely the pixel tile information with at least one of a plurality of templates, wherein the match is determined from looseness intervals between the templates and the pixel tile information. Then, outputting a portion of enhanced pixel tile information wherein the enhanced pixel tile information is formed from data associated with a matching template. The output signals are such that a preferred thresholding of the image is performed.

Abstract: A method for neutral pixel detection using color space feature vectors wherein one color space coordinate represents lightness is provided. The method includes the following steps: a) receiving an input image represented in a first color space; b) converting the input image to a second color space wherein one coordinate represents lightness; c) selecting a pixel in the second color space representation to be classified; d) computing second color space feature vectors associated with the selected pixel; and e) classifying the selected pixel between neutral and color classes based on the values computed for the second color space feature vectors. Typically, the input image is processed using a smoothing filter to create a smoothed input image prior to the conversion to the second color space. The method can be adapted to page processing or strip processing schemes with respect to the input image.

Abstract: A document processing system including a plurality of model images stored in a memory is provided. The model images are represented by a first information set with the first information set varying as a function of object-based coordinates. At least one of the plurality of model images includes a text containing region having index information intended for use in storing one or more document pages. Moreover, a query image represented as a second set of information varying as a function of object-based coordinates is provided. In practice, an image localization module corresponds the second set of information with a portion of the first set of information to obtain the text containing region, and a text extraction module extracts the index information from the text containing region to facilitate the storing of the one or more document pages.

Abstract: A method of color transformation is based on non-separable tiling of a color transformation lookup table. Non-separable tiling allows the lookup table to have fine granularity where it is needed or desired and coarse granularity where it can be tolerated. Therefore, non-separable tiling of the lookup table provides improved transformation quality for a given lookup table size. The non-separable tiles are enumerated. An appropriate tile is selected for use in a color transformation by finding an intersection between lists of tiles associated with each value that comprises a particular pixel. In one embodiment, a tile or transformation rule is associated with bit position in a bit vector. Any particular parameter value may be associated with one or more tiles or transformation rules. Therefore, a particular bit vector associated with a particular parameter value may have a plurality of set bits.

Abstract: What is presented is a method for feature recognition using loose-gray-scale template matching including at least an initial point for locating within the received image a plurality of pixel points in gray-scale surrounding an initial point. The method has the steps of first locating the initial point and the plurality of pixel points to define feature boundaries and then generating a looseness interval about the located initial point with template information being associated therewith. The next step involves determining which one of a plurality of templates for fitting fits within a threshold looseness interval, and then outputting a signal associated with that recognized feature.

Abstract: A process for color graphics image processing, related to detection and segmentation of sweeps, is provided. An input graphics image is transformed into a three-dimensional histogram in an appropriate color space 104 (e.g., CIELUV). Two-dimensional histograms are estimated from the three-dimensional histogram 106. The two-dimensional histograms are processed to detect and segment sweeps 108. Sweep segment information from the processing of the two-dimensional histograms is combined 110. The combined sweep segment information is used to process the input graphics image to identify and segment sweeps 112. Post-processing may be optionally and selectively used to reject false alarms (i.e., areas falsely identified as sweeps) 114.

Abstract: A method for implementing a filter on a signal is disclosed wherein interval membership information is computed and stored in such a manner so as to use a minimal amount of memory thereby allowing filter computation by a small number of deterministic sequence of table lookups and bit-wise logical operations. In general, the present invention involves using a non-linear filter represented as a plurality of intervals and the discrete values a sample may take. Each sample corresponds to a component in a vector. Each interval is comprised of a lower and upper vector. Each sample value is compared to the lower and upper values of the sample's associated coordinate. A table for that sample is then constructed with binary entries where a ‘1’ denotes that a sample value is within the lower and upper values of the coordinate and ‘0’ denoting otherwise. A table is built for each sample.

Abstract: What is presented is a method for rendering gray half-toned images from a received image wherein the received image is bitmap data including at least a plurality of gray-scale pixel tiles that define the image. The method includes the steps of first receiving a gray image so as to extract pixel tile information of the received image. The next step is matching loosely the pixel tile information with at least one of a plurality of templates, wherein the match is determined from looseness intervals between the templates and the pixel tile information. Then, outputting a portion of enhanced pixel tile information wherein the enhanced pixel tile information is formed from data associated with a matching template. The output signals are such that a preferred thresholding of the image is performed.

Abstract: A Document Image Segmentation method is disclosed for matching a plurality of templates with a received image wherein the received image being bitmap data including at least a plurality of gray-scale pixel tiles that define the received image, the matching method having the steps of first receiving said image data so as to extract pixel tile information of said received image wherein said pixel tile information being of a predetermined matrix size; then matching loosely said pixel tile information with at least one of a plurality of templates so as to generate pixel-wise looseness interval values there between; and finally outputting an identifier associated with the matching template such that said identifier indicates a classification. The classification is preferably based on at least one of continuous tone pictorials, text, half tones, high/low frequency range; and line art graphic. Also, preferably the classification contains information about the image source.

Abstract: A method for resolution conversion for re-sampling anti-aliased images is disclosed which decreases bandwidth costs associated with anti-aliased line art and other costs associated with interpolating these images to a desired resolution. The present method first involving the receipt of an image which is comprised of bitmap data including at least a plurality of gray-scale pixel tiles that define the image. Then receiving the image data at a first resolution and extracting pixel tile information of the received image at a second resolution. The method has the step of next using loose gray scale template matching on each of the pixel tile information with at least one of a plurality of templates so as to generate pixel-wise looseness interval values there between. Then, outputting a portion of the enhanced pixel tile information wherein the enhanced pixel tile information is formed by a matching of a template with pixel-wise looseness values.

Abstract: A method and system for image processing, in conjunction with classification of images between natural pictures and synthetic graphics, using SGLD texture (e.g., variance, bias, skewness, and fitness), color discreteness (e.g., R_L, R_U, and R_V normalized histograms), or edge features (e.g., pixels per detected edge, horizontal edges, and vertical edges) is provided. In another embodiment, a picture/graphics classifier using combinations of SGLD texture, color discreteness, and edge features is provided. In still another embodiment, a “soft” image classifier using combinations of two (2) or more SGLD texture, color discreteness, and edge features is provided. The “soft” classifier uses image features to classify areas of an input image in picture, graphics, or fuzzy classes.

Abstract: A method and system for detecting registration markers having the shape of an `X` on a scanned test target used to calibrate an image processing system. The procedure is impervious to scanning and printing distortion. A captured image is converted to a binary image and then segmented into a list of eight-connected components. A set of features is extracted from each component in the list. An accurate decision rule is executed on each component. If a component is found to be a marker, its centroid is returned as its position.

Abstract: A method and system automatically determines the width (i.e., left and right edges) of a document when it is loaded for scanning or copying. When a document is staged, the scanner collects a predetermined number scanlines of the leadedge of the document. Both first order and second order statistics of the scanner and document are then calculated based on the scanned data. The auto-width detection process is adaptive to CCD sensor output variation, document variation, and scanner variation, thus making it robust against dust in the sensor path and random fluctuations of the scanner.

Abstract: A method for generating forms using halftones along with a set of low-complexity image processing steps to extract the characters for recognition. The method exploits texture differences between character strokes and halftoned boxes or text fields. Form frames are rendered as black and white halftones and characters are extracted by exploiting differences in texture between the frames and the character strokes. A sequence of simple image processing operations, easily done in hardware, eliminates the frames while leaving the characters intact. Halftones, giving the appearance of a light color as in the dropout-color method, are easily produced by page description languages; thus, blank and filled-in forms can be scanned, printed, stored and photocopied at low cost.

Abstract: The present invention handles fully-lined, semi-lined and line-less cell tables by identifying the cells and cell separators during page recomposition processes as part of optical character recognition processes. The invention accomplishes such by iteratively identifying cell separators and cells. The processes accomplishes this by iteratively merging word boxes into cells, iteratively finding separators, and iteratively merging cells bounded by the same separators, and repeating these steps until the correct cell structure is found. With this method, rows are estimated, close words are merged into cells, columns are then estimated, cells within columns are merged, columns re-estimated, cells in the same row and column are merged into bigger cells, and then rows and cells are merged according to the detection of various table styles. This invention handles large complex tables with multiple lines of symbols per cell. This method handles multiple line cells in lined, semi-lined and line-less tables.