Abstract: A document image scanner; a camera having a field of view, for sensing differentiated information areas on each physical document as the physical document is transported; an image processing system associated with the camera, for producing a virtual image data file from the differentiated information areas sensed by the camera; sensors for determining the position of the physical document relative to the camera field of view; of electro-mechanical devices for actuating the transporter; and a data interface controller for receiving data from the sensors and controlling the electro-mechanical devices. A communications link is provided between the data interface controller and the image processing system for synchronization by the data interface controller, of the physical document as conveyed and the virtual document image captured by the image processing system.

Abstract: In a broad aspect of the invention, the master printed form is scanned and processed according to a forms definition program to produce a virtual form file comprising a virtual layout of the significant regions such as bubble targets of printed material on the form, on a virtual coordinate system. Production forms that have been marked by subjects (e.g., students, voters, survey respondents, etc.), are then scanned to produce a marked form file of gray scale darkness values for each marked form. The marked form file and the virtual form file, are compared and processed to determine the location and spatial relationships of the marks on the marked form, in relation to the virtual coordinate system of the virtual form. The raw scan of each marked form is also processed to determine whether darkened areas on the marked form image should be interpreted as intentional responses from the subject, at the virtual coordinates where targets are located on the virtual form.

Abstract: The presence of overlapped sheets on a paper transport is detected by employing such sheets as an acoustic interference filter. A beam of ultrasonic energy of appropriate frequency, angularly oriented to the planar transport path, will be attenuated to a much greater extent than the attenuation calculated based on the attenuation of a single sheet, as a result of destructive combining of wavefronts reflected from the facing surfaces of the overlapped sheets.

Abstract: The present invention includes methods of correcting optical character recognition errors occurring during recognition of alphanumeric character strings contained within one or more predetermined types of alphanumeric character fields. The methods may be practiced with a document processing system having (1) a optical character recognition device for scanning documents and outputting bit-map image data; (2) a recognition engine for converting the bit-map image data into possibly correct alphanumeric characters with associated confidence values; and (3) at least one lexicon of character strings consisting of a list of at least a portion of all of the possible character string values for each of the fields being processed. The present invention corrects OCR errors by performing a contextual comparison analysis between the alphanumeric characters outputted from the recognition engine and the lexicon of character strings.

Abstract: A block of printed information, for example an address, is located on a moving page and data commensurate with the block only is transferred to downstream processing apparatus such as an optical character reader. The location of the block of information is accomplished by generating a real time binary profile of a data stream produced during scanning of the page and comparing the binary profile with a plurality of prerecorded profiles commensurate with the type of information to be located.

Abstract: Digital matrices, for example digitized images, are enhanced by comparing a matrix to be enhanced with plural enhancement patterns for each position in the matrix of each element of the matrix to produce pattern codes commensurate with the matches with the enhancement patterns. The pattern codes are employed to generate plural enhancement decisions for each element of the matrix and these enhancement decisions are combined to provide a single enhancement action for each element of the matrix.

Abstract: The processing of documents evidencing business transactions is enhanced by optically scanning and imaging only specific areas of interest, amount fields for example, and simultaneously storing the imaged amount fields in two separate files. Subsequently, only those images which correspond to a dollar amount which has not been successfully read will initially be displayed to an operator for key entry of the amount. A balancing procedure is performed and if a comparison of the total of the individual imaged amounts equals the total entered on a summary document further key entry of amounts will be unnecessary.

Abstract: The recognition of patterns is accomplished through boundary tracing and subsequent storage of encoded range testable data commensurate with the occurrence, interrelationship and orientation of geometric features on the pattern boundaries. The encoded data is compared with generalized prototypes which define the geometric shape of all probable permutations of all possible patterns to be recognized, a match between the encoded data and a prototype constituting identification of the pattern.

Abstract: A stacking and feeding device of document-handling equipment for documents moving along one of their edges, includes a transporting conveyor belt system that forces the documents against a stop where they begin to stack up in the horizontal direction in the order in which they arrive. A feeder rides on the outermost document of the stack and slides along guide shafts as the size of the stack changes. The feeder is selectively actuated to remove the documents from the stack one at a time in the order in which they entered the stack. A swing arm is pivotally and extensibly connected between the feeder and the following stationary part of the equipment so that the documents can be delivered thereto, regardless of the position of the feeder.

Abstract: A document transport for a document scanning device has an elevator for locating a stack of documents to be fed automatically into position to be scanned. The top document is removed from the stack by a feed roller. Other documents which may become entrained with the top document are returned to the stack by separation rollers. The top document is driven into the scanning device against an alignment gate and this action eliminates any skew in the position of the document. At a read or scan station beyond the gate, which is reached by the document once the alignment gate has been withdrawn, the data on the document is scanned as the document is moved by read rollers in a controlled fashion. As the document leaves the read station it is trapped between the two single opposing belts in a buffer unit, which move the document at the spped of the scanning rollers.

Abstract: In a character recognition system lines of data are stored in a line memory. This data is then accessed by a hardware implemented character position detector over a special line memory bus, i.e., a LM-bus. The detector searches the memory for the margins of characters and can also perform a count of the black points or character points within the margins. The position detector is activated by a system controller over a B-bus, but once started by the controller it will run on autonomously until it has completed its task, thereby leaving the processor and the B-bus available for other functions.

Abstract: In a character recognition system lines of data are stored in a line memory via a series of multiplexers. The data can originate in a controller for the system, a character reading station or other sub-systems, and is supplied to the line memory over a controller bus, a transport bus and a line memory bus, respectively. Although stored in line form, two-dimensional sub-arrays of the data in various configurations can be accessed directly by the other sub-systems of the recognition system via the multiplexers and the line memory bus, without the interaction of the controller bus.

Abstract: As described herein, a program controlled image dissector tube scans the printed information recorded on a storage medium in a first raster to provide analog information signals representative of segments of the recorded information and in a second raster to provide analog information signals representative of all the recorded information. If, as determined by the analog information signals representative of segments of the recorded information, the printed information is legible, the analog information signals representative of all the printed information are converted into digital form and scaled, if necessary, to reduce the digital character data to a predetermined scan length. Selected arrays of the digital character data are then scanned to develop black or white signals about a center bit within such arrays. The black or white signals are, in turn, accumulated to provide an array of black and white signals representative of an entire character.