Abstract:
Characters are located in a character recognition system (10, 12, 14, 15) by providing images in response to scanning characters, and then smearing black pixels in an image in up and down and left and right directions for detecting corners which are used to produce candidate character frames. These candidate character frames are scored based on how well they correspond to character size and expected location to determine which character frames are to be submitted for character recognition.
Abstract:
A leak tolerant liquid cooling system for electrical components (10) comprises a cooling circuit which contains a pump (17) and conduits (14, 16, 18, 20) that circulate a liquid coolant (LC) past the electrical components (10) and through the bottom chamber (15b) of a purge tank (15). This purge tank (15) also has a top chamber (15a) which is connected to the bottom chamber (15b) through a passageway (15d); and, the bottom chamber (15b) is sized such that the liquid coolant (LC) passes through it with a velocity which is low enough to let any air bubbles in the coolant rise and move by buoyancy through the passageway (15d) into the top chamber (15a). Air is purged from the top chamber (15a) through a valved output port (25) by forcing liquid coolant (LC) into the top chamber (15a) through a valved input port (24). To complete this purge quickly, the passageway (15d) between the purge tank chambers (15a and 15b) is configured in a way which hinders the flow of coolant (LC) from the top chamber (15a) to the bottom chamber (15b) while the coolant (LC) is being forced into the top chamber (15a). As the air purging occurs, the coolant (LC) continues to be pumped past the electrical components (10) through the cooling circuit. Since the pump (17) is not switched ON and OFF during purging, pressure surges in the cooling system do not occur, and switching noise in the voltages to the electrical components (10) is not generated.
Abstract:
Apparatus for implementing input/ouput (I/O) operations in a computer system operating under the control of a UNIXTM operating system includes a dedicated communications processor in addition to the main or host processor. These two processors communicate via a shared memory which may be independently accessed by each processor. The function performed by the terminal I/O driver and the line discipline modules are divided between the host and communications processors. The communications processor performs all canonical processing of the data received from the terminal I/O devices. It also maintains a data structure that indicates the instantaneous status of each terminal I/O device. Using this data structure, the communications processor is able to operate in a substantially interrupt-free environment, polling only those I/O devices that are indicated, in the status data structure, as needing service. A message facility in the shared memory controls communication between the processors. Each processor sends messages to the other through a dedicated circular queue. To receive a message, a processor examines the sending queue of the other processor. Response messages are sent back to the sending processor using the same message buffer as was used for the original message.
Abstract:
The message management system utilizes a flat file (20) on mirrored disk for storing fixed-length Message Segments (MS) of variable-length Messages to be managed by the system. Inconsistencies between mirrored copies of data are tolerated and corrected. Message Management includes Storing, Retrieving and Deleting Messages in response to commands from clients. Each Message Segment in the flat file is stored in a record containing the Message Segment data and Recovery Information (RI). A data base (21) for accessing the Message Segments in the flat file is maintained on system disk. The Recovery Information is such as to permit recovery of the data base from the Recovery Information in the flat file if the data base is obsolete, lost or corrupted. TNAUDIT file is utilized to determine that the data base is synchronized with the flat file. If data base inconsistency is detected, the data base is rebuilt from the Recovery Information.
Abstract:
An apparatus for recognising characters (20) on a document (10) comprises a scanning head (24) and decoding circuits (78, 72, 86, 92, 80), feeding a controller (84). The controller (84) detects the identity of the characters (20A - 20N) by calculating the ratio found by dividing the time between a first positive pulse (32) and a second positive pulse (34) by the time between the first positive pulse (32) and a last negative pulse (36) and comparing it against a stored table of ratios.
Abstract:
An image processing station (18) displays data on a monitor (22) of images captured by an image capture work station, compressed by data compressor (12), stored in an image store (16), and retrieved. The image processor station can accept commands from a keyboard (26) or output data and images via a printer (24). A decompressor (36) decompresses retrieved data from the store (16) and a video memory (50) comprises selectable first (56) and second (62) pages each of which may be loaded with fresh image data whilst the other is being used for display. Images, on successive screens, are swapped from a first position to a second position whilst fresh images are moved into the first position.
Abstract:
Check processing apparatus including an offset pair of image-lift sites and an associated ''Heat-Dump'', with this apparatus preferably built as an add-on, plug-in module to be used in a check-sorting system. A lamp projecting a high-intensity optical beam of spectral radiation, with an IR-extracting (cooling) array including an IR-dissipating ''black-box'' arranged to receive IR and convert it to box-heat, plus an IR-diverting ''hot-mirror'' interposed along the beam to divert (only) the IR to this ''black-box''. An optical array including a source projecting spectral light onto the input-aperture of a fibre-optic bundle, the bundles exit-aperture outputting a rectangular beam with fibres ''random-laid'' and spread-along the exit-aperture to make this rectangular beam's intensity more uniform. A check Sorter-Image-Processor with two (or more) imaging sites, each lit by a beam through a fibre-optic array, with a lens focusing the beam on each array such that a shift in lens-focus reduces output light from the array, the two sites being thus ''matched'' in light-intensity by merely focus-adjusting these lenses. Technique of integrating cooperating optical-thermal components by providing a single casting for mounting, and positioning them under the working platform of an operating structure, with the casting adapted to be simply, removably coupled to this platform, and arranged to thereby thermally-optically isolate these components from interference with or by this operating structure. Arrangement for lifting image of checks, with Xenon lamp, Fluorescent lamp or like source of stable-wavelength-output and CCPD Camera means selected to exhibit ''close-to-human'' spectral response plus a photopic filter, disposed upstream of the Camera and adapted to shift its received image-light spectrum in the direction of that which is optimal for the human-eye.
Abstract:
Disclosed are power encoder means for imprinting MICR characters on checks, with optical check-sensing means disposed along a check-transport path, including optical skew-sensor means. Print drums are described wherein a special ''alignment character'' is placed, along with print-symbols, on the surface of the print drum, mixed-in with at least one of the character sets. Disclosed are impact printing arrangements with impact surface means, wherein impact-energy is graded and applied according to the die-surface area of a selected die-symbol. Disclosed are impact printing arrangements, with multi-column die-surface means and an associated plural print-hammer means, one for each column, wherein techniques are taught for monitoring hammer motion and coordinating simultaneous impact thereof on said surface means by detecting and analyzing hammer-activating inductor voltage. Disclosed are servo methods and apparatus whereby unit-record documents are transported past a process-station according to one, or several, ''velocity-profiles'', depending upon document type and/or condition. Disclosed are ''anti-wander'' web-transport arrangements wherein web-means are intermittently advanced, while being monitored for misalignment along the transport-path, and also being automatically urged back into proper alignment along this path.
Abstract:
A liquid cooled integrated circuit module (22) includes a substrate (10), a plurality of chips (11) mounted on the substrate, and electrical conductors (12) integrated into the substrate to interconnect the chips. A compliant member (20) which is completely seamless overlies all of the chips. This seamless compliant member is hermetically sealed at its perimeter to the substrate around all of the chips. Between this seamless compliant member and the chips are thermally conductive studs (21), and they carry heat by conduction from the chips to the compliant member. A rigid cover (22) overlies the compliant member, and it is attached to the substrate at its perimeter. Within the cover are several parallel spaced apart ribs (22A) which project towards and press against the compliant member between the chips, and they form channels for a liquid coolant which carries heat away from the compliant member.
Abstract:
A data compression/decompression apparatus employs common circuitry and a single string table for compression and decompression. A throttle control is provided to prevent data under-runs and an optimizing start-up control delays the start-up of the recording device (114) until the compression apparatus has compressed sufficient data to efficiently reduce throttling and loss of compression when the output device is started. The decompression apparatus may operate to decompress compressed data when the compressed data is read in either the same direction as it was recorded, or read in the direction reverse to that in which it was recorded (116). A further feature is the provision of a counter which is incremented by one after a predetermined number of string codes have been written into the string table. The output of the counter is stored in the string table with each string code and prefix code. When searching for an empty or usable location in the string table, the count value read from the location is compared with the count in the counter. If the two counts are not equal then the location is considered ''empty'' and may be written into. This arrangement avoids the usual procedure of intermittently clearing each location of the string table individually since stepping the counter is equivalent to clearing the entire table.