 |
Description  |
|
|
BACKGROUND OF THE INVENTION
This invention is directed to information storage and retrieval, and more
particularly to a system for storage and retrieval of large quantities of
documents which may include text, illustrations or combinations thereof.
The invention is particularly useful in the archival storage of historical
documents wherein it is desirable to maintain the integrity of the
historical document, including its original appearance. As used herein and
in the appended claims, the term "archival document shall be used to refer
to a document containing textual information but wherein the appearance of
the original document, and not merely the content of the text, is
significant, and an "archival document storage system" or "archival
document image storage and retrieval system" shall be used to refer to a
system for storing and retrieving images of archival documents where the
appearance of the original document is of interest to the user of the
system.
In a conventional archival storage system, documents would typically be
separated into files. In some cases, large quantities of documents would
be stored with no practical indexing, so that reasonable access to the
documents would not be available to any one but a few scholars who knew
where to find them. An indexing system could be manually generated and
researchers could access the stored documents through an index card file,
but manual indexing systems have not proven entirely satisfactory.
When using an index card file system, it is still necessary to retrieve a
document from the storage files in order to determine if it is relevant.
If relevant, it is then necessary to obtain a photocopy or other
reproduction of the document. These processes can take considerable time
where a large number of documents are involved.
The handling of the documents contributes to the deterioration of the
documents, which can be a long term problem in an archival storage system.
Still further, the complexity of manually generating an index card file
system can itself be a disincentive for maintaining such a system when
extremely large numbers of documents are involved.
It is desirable to provide some type of automated search capability, and it
is known in some systems to index documents, e.g., by key words, and to
permit automated searching. However, this facilitates only the searching
aspect of the conventional system described above, and it is necessary to
manually retrieve documents, to take the document to a photocopy station
to obtain a copy of the document, and to manually generate the key words
and phrases which will be used in the indexing system.
In at least one publicly available storage and retrieval system, i.e., the
automated search system currently maintained by the U.S. Patent and
Trademark Office for searching U.S. patents related to data processing,
pertinent portions of the documents covered by the data base are stored on
microfiche. Each document must be read by Patent Office personnel who will
then assign that document to one or more descriptive headings. A system
user can then key in a particular heading or a plurality of headings
combined with logical operators, and the system will display from
microfiche the stored portions of every document satisfying the search
request. While such a system represents a substantial improvement over
entirely manual systems, it is still not entirely satisfactory in a
document storage and retrieval system employing very large numbers of
documents, e.g., many millions of pages of text and drawing. The
microfiche storage capacity is insufficient for such large numbers of
documents, and the speed of retrieving the appropriate microfiche for
display would also be unsatisfactory in a system of great size. Further,
the pages of documentation are recorded on microfiche by a conventional
photographic process, and there is no opportunity for the system to
recognize the content of the documents being photographed. All key words
and descriptive headings must therefore be manually entered. Still
further, it is necessary for each viewing station to have its own set of
microfiche, or at least for all viewing stations to be located immediately
adjacent the microfiche file.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a storage and retrieval
system which overcomes the above-described drawbacks.
It is a further object of this invention to provide such a system which is
capable of storing very large numbers of documents in an efficient manner.
It is a still further object of this invention to provide such a system
which will permit rapid automated searching of the stored documents,
retrieval of the documents for display, and on-site printing of the
documents while maintaining the integrity of the appearance of the
original document.
It is yet a further object of this invention to provide a document storage
and retrieval system wherein the requirement for manual data entry is
minimized to substantially ease the burden of generating the desired
document index.
It is a further object of this invention to provide a document storage and
retrieval system wherein a remote user can access the data base for
storage, retrieval and printing of documents including both printed text
and drawings.
Briefly, these and other objects of the present invention are achieved by
an archival mass storage and retrieval system using a digital camera to
capture data in a digital form. The output of the digital camera is
provided as an input to a general purpose digital computer which, in turn,
is provided with an optical disk storage system. The digital information
representing the object scanned by the digital camera is stored for
subsequent retrieval on the optical disk together with appropriate index
information.
The digital camera can thus be used to scan photographs, blueprints, pages
of books, memos, etc., or even three-dimensional objects, and can
accurately transform any of these into data which can be subsequently
handled by the computer system. With all data in digital form, the
documents can be transmitted to or from remote locations. The computer is
preferably also provided with software necessary for recognizing text
contained in the bit-map output of the digital camera so that, upon
recognizing the content of any document, the appropriate index information
can be automatically generated. The captured images and associated index
information are then stored on an optical disk for future searching and
retrieval. Retrieved documents can be displayed on the computer monitor
and can be reproduced on any commercially available printer having
graphics capability.
BRIEF DESCRIPTION OF THE DRAWING
The invention will be more clearly understood from the following
description in conjunction with the accompanying drawing, wherein the FIG.
1 of the drawing is a block diagram of essential components of the storage
and retrieval system according to the present invention, and FIG. 2 is a
brief flow chart illustrating an efficient procedure for document capture
and indexing.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
A block diagram of the overall system of the invention is shown in the
accompanying Figure. The system includes a digital camera 10 for scanning
documents and converting the image information of each document to a
digital information stream. The digital output of the camera 10 is
provided to a processor 12, which may be any one of a number of suitable
general purpose digital computers. The processor 12 receives the image
information and preferably stores this information in a buffer storage,
e.g., disk storage device 14. The processor 12 then generates the
appropriate indexing information, with the help of additional data
manually entered via keyboard 20, if necessary, and the multiple frames of
image information together with the associated index information are
stored on a digital storage medium in a manner to be described in more
detail below. In a preferred embodiment of the invention, the digital
storage medium will be an optical disk, although other types of high
density storage devices would be acceptable. The index stored on the
optical disk can then be searched by the processor 12 to obtain the
addresses of frames having requested information. The disk locations can
then be accessed to retrieve the video information for display on a
display 16. If desired, the displayed video information can be reproduced
via printer 18.
The individual components and their interaction will now be described in
more detail.
With regard first to the digital camera 10, the camera should be a high
resolution digital camera, such as a scanning-type CCD camera having a
resolution of on the order of 3000.times.1500 pixels. One example of a
suitable camera for use in this invention is the Model 610 camera
available from DataCopy of Mountain View, California. The camera is
available with a suitable software routine to permit the processor 12 to
receive and process the digital output from the camera for either storage,
display, printing or retransmission, e.g., to an optical disk recorder.
The video information provided by the camera 10 may represent a typewritten
or handwritten document, a blueprint, photograph or even a physical
object, and the output of the camera 10 will be a digital bit stream
provided to the processor 12. The processor 12 may be any one of a number
suitable general purpose digital computers, such as the PC XT
microcomputer available from International Business Machines Corporation.
The processor stores the digital video information in a disk buffer
storage device 19. In the case of a typewritten document, the processor
may be provided with software for converting the bit-mapped video
information into ASCII character data, so that the processor can determine
the information content of the text. Having determined the information
content of the text, the processor employs appropriate software to
generate the index and cross referencing information which can also be
stored in buffer 14. The software should preferably be a self-index
software program which will generate a data base index from full text, so
that every word in the text will become a key word in the index. A
suitable indexing software program would be the ZyINDEX program available
from ZyLAB Corp. of Chicago, Illinois.
The text recognition software for recognizing text and generating ASCII
characters from bit mapped image data is presently available and/or can be
generated from commercially available software in a very straightforward
manner. For example, commercially available OCR software in effect
examines the bit map, or pattern of pixels, of the character image as a
scanner moves across a page. Thus, the scanner momentarily "captures" the
image of each character and recognizes each captured character image
before going on to the next character. To recognize in software the
characters represented by a captured digital image of an entire document,
it is merely necessary to scan the digital image in the same manner as the
original document would have been optically scanned by the OCR device. For
example, the captured image could be displayed on a monitor, and a cursor,
which may preferably cover substantially the same area as would be covered
by an OCR scanner, may be moved across the monitor screen simulating the
same scanning motion as an OCR scanner. The image portion covered by the
cursor would be recognized by the OCR software in the same manner as is
conventionally done. One example of OCR software suitable for modification
for use in the present invention is the software used in the OMNI-READER
optical character reader available from Oberon International having
offices in Irving, Texas and London, England. The only modification
necessary to this software would be to substitute the digital image
information, from a region of the image covered by a cursor, for the image
information signal which would normally be provided to the software from
the scanner output.
Alternatively, software for recognizing text from bit mapped image data is
available from Bell Northern Research, P.O. Box 3511, Station C, Ottawa,
Canada.
If it is desired to utilize in the indexing routine additional key words
for a particular document which are not found in the recognized text, such
additional key words, e.g., category headings, titles, etc., can be
entered via keyboard 20. For documents which are drawings or other images
which cannot be recognized by the processor 12 or if the document is a
typewritten document which has deteriorated to such an extent which
precludes machine recognition, appropriate key words and/or titles will be
entered via keyboard 20 for use in the indexing routine.
As described above, the preferred embodiment of this invention will use an
optical disk storage, and the invention will be described in this context.
However, it should be appreciated that other storage devices could be
used, and that the invention is not limited to use with optical disk
storage but is instead limited only by language of the appended claims.
The least expensive technique for optical disk recording is to generate a
master disk from 1" C tape, with the image information being represented
by analog signals. Accordingly, the digital information from the buffer
storage 14 can be provided in the form of digital signal to a translator
22, e.g., such as available from Matrox Electronics Systems, Ltd. of
Montreal, Canada, which preferably performs a simple D/A conversion to
provide an analog video signal at its output. The analog video signal
output from the translator 22 can then be recorded on standard video tape
24, e.g., of the type well known in the art. The indexing information
would then be similarly provided through the translator 16 for recording
on the tape 24. The order in which the video information and index
information are provided for recordation on the tape 24 could be reversed,
but it is preferable that all of one type of nformation be recorded
followed by the other type rather than interleaving video and index data.
The tape would then be used to generate a master disk in a disk mastering
station 26, e.g., available from Phillips. With the video and indexing
information recorded on different areas of the disk, it is a simple matter
to designate certain track numbers as index storage area to facilitate
accessing of the index information.
Since the index information is in ASCII format rather than bit-mapped
image, it would also be possible to record the index information in
digital form on the tape 24 and master disk. In such a case, the index
information could be provided through the translator 22 without
conversion, and the separation between image and index information on the
optical disk could be indicated merely by noting the track number above or
below which a digital signal format is employed.
After generating a desired number of disk copies, e.g., plural copies for
distribution to remote accessing stations, a recorded optical disk may be
placed in an optical disk playback device 28. This may, for example,
comprise a standard laser disk player commercially available from several
sources. The laser disk player could have an RS-232 or other suitable
computer interface for coupling to the processor 12, and the disk player
28 can then be accessed and controlled by the processor 12. For search and
retrieval, the operator would enter via keyboard 20 a key word to be
searched, and the processor would then search the index information on the
optical storage media to determine the disk location of any documents
responsive to the key word inquiry. These locations would be sent back to
the processor 12, which would then control the disk player to access each
of the locations and retrieve the image data. The bit-mapped digital image
data could then be displayed on a display (e.g., a monitor) sufficiently
high resolution to display not only retrieved text but also detailed
picture images, e.g., a vertical resolution of not less than approximately
1500 lines. Such monitors are available from a number of sources, e.g.,
DataCopy. If any of the retrieved documents are considered sufficiently
pertinent, a copy thereof, either text or video image, or both, can be
reproduced on a typical graphics-quality printer 18.
In one example of an application of the system of this invention, a large
quantity of documents would be successively "photographed" by the camera
10, and the digital information would be stored in buffer storage 14.
After storing a sufficient quantity of image data in the buffer storage
14, the processor 12 would provide the digital data for recording on high
density storage media, e.g., either through translator 22 and tape 24 to
the optical disk mastering station 26 or directly to a digital optical
disk recorder. If documents are to be added to the system at a relatively
slow rate, it may be satisfactory to perform the text recognition
substantially simultaneously with the capturing of the image information
by the digital camera 10. However, if large numbers of documents are to be
captured and indexed, the most efficient procedure would be to
successively capture the documents and to generate an optical disk, or
indeed a plurality of optical disks, storing only the captured image
information from the documents. After storing the digital image
information, the images can be retrieved from the disk players and the
index information can be generated at that time either by text recognition
software or by entry of index information via keyboard 20, and the image
data together with the corresponding index information can then be stored
on a new set of optical disks. This operation is briefly illustrated in
the flow chart of FIG. 2.
Thus, since the digital image information can be recognized later just as
easily as at the time of original capture, the image capturing process
need not be delayed while the indexing software and text recognition
software are operating. Indeed, with the ability to recognize and index
the captured images at a later time, it is possible to capture all of the
document images on one processor while performing text recognition and
indexing on another processor, thus subtantially enhancing the throughput
speed.
Since constant changing of the optical disks may be impractical, the
storage capacity may be increased either by using a multiple-disk player
or by stacking a plurality of disk players and separately addressing each
player. A suitable multiple-disk player is available from Mitomo
Corporation and provides an expanded storage capacity of on the order of
50 individual disks. However, accessing speed would be improved by
utilizing a plurality of separate players. Due to the present low cost of
disk players, the second alternative may be preferable.
It should also be noted that, in either type of system employing multiple
disks, it may be preferable to dedicate entire disks to index information
rather than particular sections of each disk.
With regard to the optical disk recording, the embodiment of the invention
reduced to practice employs the translator 22, 1" C tape 24 and disk
mastering station 26. This requires that the translator perform D/A
conversion when providing the information from the processor 12 to the
tape 24, and it also requires that the translator 22 perform A/D
conversion when image information is read from the disk player 28 and
provided to the processor 12. However, optical disk recording technology
has advanced to the point where the image data and index information could
be recorded directly rather than via the tape 24 and disk mastering
station 26. This would also permit elimination of the translator 22.
Instead, the information from the processor 12 could be provided directly
to a digital optical disk recorder 30, recorded in digital format, and
played back through a digital disk player 32. However, present optical
disk recorders are somewhat expensive and may not yet be justifiable from
a cost standpoint.
In the preferred embodiment described above, the processor 12 is provided
with appropriate software to generate ASCII code from bit-mapped image
data, so that the indexing information could be automatically generated
with minimal operator input. In those cases, such as drawings or
unreadable text, where the processor 12 cannot generate its own indexing
information, the necessary information could be entered via keyboard 20.
In the case of unreadable text, it would normally be sufficient merely to
enter a short description of the documents being recorded. However, in
those instances where it is desirable to enter a substantial portion or
all of the text, it may be preferable to have the text transcribed at some
other clerical station and entered into the processor 12 via an Optical
Character Recognizer (OCR) 34. Indeed, if the processor 12 is not provided
with software for recognizing the content of the images provided from the
camera 10, an OCR 34, e.g., of the type available from Kurzweil Computer
Products, Inc., 185 Albany Steet, Cambridge Massachusetts, or DEST
Corporation, 1201 Cadillac Court, Milpitas, California, can be used to
read and recognize all or a substantial portion of the documents being
stored.
It should be noted that the OCR is not used as a substitute for the digital
camera, but rather only as a substitute for recognition software in the
processor 12 which would otherwise be used to recognize the textual
content of the digital camera output. In other words, the digital camera
10 is used in all instances to generate digital image information
representing the document and this digital image information is eventually
stored on the optical disk or other high density storage medium. The
information generated by the recognition software in the processor 12, by
the OCR 34, or by manual entry via keyboard 20, is used solely for
indexing purposes.
For large drawings, e.g., blueprints, it would be possible to enter the
image data via a digitizer/plotter 36 providing a digital bit-map output
corresponding to the drawing. A suitable digitizer/plotter is available
from XYZ TEK Corp., of Englewood, Colorado. The digitizer/plotter 36 could
also be used to reproduce copies of blueprints and large drawings which
may be impractical for reproduction on the printer 18.
A modem 38 may be coupled to the processor 12 for bidirectional
communications. For example, image and/or index data could be generated at
a remote location and transmitted to the processor 12 in a well-known
manner. Further, the modem would allow the data base to be accessed from
remote locations by researchers and scholars around the world. In response
to a query received from a remote location, a particular document, which
may be text or a drawing, could be sent via modem for remote display or
remote printing.
System throughput may be enhanced using commercially available automated
document positioning systems to permit rapid scanning of successive
documents, in which case it would also be desirable to equip the camera 10
with an autofocus capability available from the camera manufacturer.
The system may also be enchanced by providing one or more portable data
capture systems comprising a camera, a portable processor having a
removable hard disk, and a simple display such as a flat panel plasma
display. The image data could then be captured and transferred at a later
date to the optical disk storage.
The document storage and retrieval system described above provides
enchanced storage capacity while simplifying the indexing and improving
accessibility to the stored information. The use of the digital camera 10
results in digitizing image information, making it possible to process the
image information for recognition of the content, transmit both text and
drawings to remote locations and print the documents on available
graphics-quality printers. Most importantly, the system provides for
automated storage and retrieval of documents with electronic reproduction
of the documents in their original form and appearance, which renders the
system highly advantageous in the archival storage of large numbers of
historical documents.
It should be appreciated that various changes and modifications can be made
to the specific embodiment disclosed above without departing from the
spirit and scope of the invention as defined in the appended claims.
* * * * *
|
|
 |
|
 |
Description |
|
