Abstract: A scheme for extracting and comparing graphs from medical reports to detect duplicates that may indicate fraudulent medical diagnostic testing results, due, e.g., to billing or insurance fraud. Domain knowledge is used to pre-process input documents and automatically extract graph images. These images are stored in a database and compared to one other using a distance-based comparison metric that is robust to noise and other image acquisition artifacts. Graphs are compared in blocks of 1000 images at a time using a two-pass comparison algorithm to identify the top matches for each graph. If a graph on the page currently being analyzed is identified as a close enough match to a known graph in the database (e.g., the graph extracted from the current patient's medical record appears to be identical to the graph of a different patient), then the page is flagged as potentially being evidence of fraudulent activity.

Abstract: An automatic character cell determining apparatus automatically determines the character cells within the text image of a document. A connected component generator means generates connected components from the pixels comprising the text image. An aligning device aligns skewed and warped lines to the proper image axes. A bounding box generator generates a bounding box surrounding each connected component. A character cell determining device for locating character cells including one or more connected components has a vertical splaying device and a horizontal splaying device for ensuring white spaces between lines and connected components, a vertical profile device for determining the vertical positions of a line, a splitting device for splitting ligatures of two or more connected components and a character cell generator for generating character cells grouping together one or more connected components.

Abstract: An automatic script determining apparatus automatically determines the gross script-type of the text image of a document. A connected component generating means generates connected components from the pixels comprising the text image. A bounding box generating means generates a bounding box surrounding each connected component. A centroid determining means determines a centroid for each bounding box. A script feature determining means determines the locations, relative to the centroid, of one or more predetermined types of features, for each bounding box. A script determining means determines a distribution of the located script features for the entire text image, and compares the determined spatial distribution to predetermined distribution for at least one script-type to determine the script type of the text image.

Abstract: A first method for exact and inexact matching of documents stored in a document database includes the step of converting the documents in the database to a compacted tokenized form. A search string or search document is then converted to the compact tokenized form and compared to determine if the test string occurs in the documents of the database or whether the documents in the database correspond to the test document. A second method for inexact matching of a test document to the documents in the database includes generating sets of one or more floating point values for each document in the database and for the test document. The sets of floating point numbers for the database are then compared to the set for the test document to determine a degree of matching. A threshold value is established and each document in the database which generates a matching value closer to the test document that the threshold is considered to be an inexact match of the test document.

Abstract: An automatic language determining apparatus automatically determines the particular Asian language of the text image of a document when the gross script-type is known to be, or is determined to be, an Asian script-type. A connected component generating means generates connected components from the pixels comprising the text image. A character cell generating means generates a character cell surrounding at least one connected component. An optical density determining means determines the optical density, in absolute numbers or percentage of pixels, of the pixels within each character cell. A script feature determining means first generates a histogram, then converts, by linear discriminate analysis, the histogram to a point in a new coordinate space.

Abstract: A method and apparatus for identifying documents and classes of documents. The documents are provided with distinctive logotypes which are preferably at the top of each document. The coding of the logotypes is by the use of distinctive angular alignments in the logotype. The logotype is scanned at different angles in order to determine angular "signatures" for comparison with a predetermined power distribution.

Abstract: An automatic character cell determining apparatus automatically determines the character cells within the text image of a document. A connected component generating means generates connected components from the pixels comprising the text image. A bounding box generating means generates a bounding box surrounding each connected component. A character cell determining means for locating character cells comprising one or more connected components comprises a vertical splaying means and a horizontal splaying means for ensuring white spaces between lines and connected components, a vertical profile means for determining the vertical positions of a line, means for splitting ligatures of two or more connected components and means for generating character cells grouping together one or more connected components.

Abstract: An automatic abstract character coding system automatically generates abstract coded characters from the text image of a document when the gross script-type is known to be, or is determined to be, a European type script. A connected component generating means generates connected components from the pixels comprising the text image. A spatial feature determining means generates a character cell surrounding one or more aligned connected component. A character-type classifying means converts the character cell to one of a plurality of abstract character codes.

Abstract: Skew angle of an image is determined based on determination of location of fiducial points on the image. Fiducial points may be located through a comparison of the scanning of a first line with scanning of a subsequent line. These fiducial points may be defined in terms of pixel color transitions located on a first scan line without a corresponding transition on the succeeding scan line. Skew angle may be determined from image data in uncompressed form or in compressed form. Where skew angle is determined from image data in compressed form, the 2-dimensional CCITT facsimile recommendations may be used. In such cases, the locations of the fiducial points may be taken as the locations of the pass codes of the compressed image data. Specifically, pass codes indicating a pass of white pixels are used.

Abstract: A database system is provided for interchanging visually faithful renderings of fully formatted electronic documents among computers having different hardware configurations and different software operating environments for representing such documents by different encoding formats and for transferring such documents utilizing different file transfer protocols. All format conversions and other activities that are involved in transferring such documents among such computers essentially are transparent to their users and require no a priori knowledge on the part of any of the users with respect to the computing and/or network environments of any of the other users. All database operations are initiated and have their progress checked by means of a remote procedure call protocol which enables client applications to obtain partial results from them relatively quickly, without having to wait for such operations to complete their work.