Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~296429
DIGITAL IMAGING FILE PROCESSING SYSTEM
BacXqround of the Invention
This invention relates generally to digital
imaging systems. More specifically, the present
invention relates to a method and apparatus for
increasing automation and document classifying
ability of such systems.
In the typical digital imaging system a specimen
(or document) is illuminated and then scanned at
every pixel to obtain analog reflectivity values.
The analog reflectivity values are converted to a
multi-digital number by an analog-to-digital
converter, where the multi-digital number ranges from
~ a minimum value representing the darkest state
:: ,
~ ,
~29642g
possible for the pixel, decimal number O for example.
to a maximum value represen~ing the lightest state
possible for the pixel, decimal number 63 for example.
The multi-digit number representing the
reflectivity of the pixel is compared to a threshold
value (sometimes dynamic), so that the final
digitized picture is represented by a single binary
digit for each pixel. Stated differently, each pixel
in the final digitized picture is stored or displayed
in either an "ON" state or an "OFF" state. In
comparing the multi-digit number to the threshold
value the resulting pixel value is obtained as
follows: if the multi-digit value exceeds the
threshold value the pixel is determined to be in the
"ON" state: if the multi-digit value equals or is
less that the threshold value the pixel is determined
to be in the "OFF" state.
SummarY of the Invention
Accordingly, an object of the present invention
is to provide means for efficiently controlling a
document scanning process with minimal user
intervention.
--2--
tl29642g
70840-105D
The foregoing and other objects of the present invention
are realized by using a "smart scanner" device, which automatical-
ly feeds a stack of input documents past its scanning apparatus,
including a charge-coupled device (CCD) array. The smart scanner
identifies highlighted data from the input documents and attaches
special significance to highlight marks of predetermined shapes.
A scanner server (software module) is employed to control the
smart scanner as well as send the image to an image file. The
image is sent to an optical character recognition (OCR) device,
controlled by its OCR server module. The OCR converts the image
data into text data and stores it in a text file~ An application
module controls the flow of image and text data as well as the
assignment of file names and memory destinations to each document.
In summary, the invention provides a digital imaging
file processing system for processing and filing digitized docu-
ments comprising: scanner means for scanning and converting a
group of input documents into digital image data; storage means for
storing digital image data; and header page means associated with
each group of documents to be digitized for automatically in-
structing the operation of the system with the use of a coded
header page that includes instructions relating to the control of
the scanner means, control of the storage means and the use, non-
use and nature of use of an optical character recognition device.
-- 3 --
1296~9
70840-105D
DescriptioD of the Drawings
~ he f~reg~ing and othe~ ob3ects and features of the
pre~ent invention are apparent from the specification, the drawings
and the two taken together. The drawings are:
Figure 1 is an is~metric drawing of an example of a
digital imaging system, including a document processor, such as
may be used in conjunction with the present invention.
Figure 2 is a block diagram of a distributed processor
version of the digital imaging file processing system in Figure 1.
; Figure 3 is a block diagram of system in Figure 2 im-
plemented with a local area network.
Figure 4 is a schematic diagram of the digital imaging
system encompassing the present invention.
Figure 5 is a sectional view of a document processor
of the same general type as is shown in Figure 1.
Figure 6 shows an example of the header or identifica-
tion page to be used with the present invention.
Figure 7 i8 a graph of charge-coupled device outputs in
millivolts of a doGument's white background and yellow highlight
marks versus the wavelength of illuminating white light.
Figure 8 is a graph of charge-coupled device outputs
in millivolts of a document's white background and yellow highlight
marks versus the wavelength of illuminating blue light.
,
~ ~ 4 ~
.
~296429
Fig~re 9 shows a digital representation of a
highlight mark.
Figure 10 shows a rectangle assimilation of the
highlight mark in Figure 9.
Figure 11 shows a highlight mark which indicates
the title of the dooument.
Figure 12 shows a sample document page with
highlight marks serving as column separators.
Detailed Description
Figure 1 shows a digital imaginq system 100 used
to create digital images o f a document 102 on a
document mount 104. The illustrated system is the
PIC system of Wang Laboratories, Inc., the present
assignee of the invention herein. The document or
specimen 102 may be a picture, chart, graph, text
paqe or other image.
The specimen 102 is scanned on a line-by-line
basis by a scanner or signal providing means 106
above the document. The primary components of the
scanner 10~ are a lamp or specimen illuminator 108,
and a camera 110.
-5-
~296429
The specimen illuminator 108 illuminates the
specimen 10`2 by causing li~ht to be projected onto
the specimen 102 in such a manner that the projected
light is reflected from the specimen 102. Opaque
documents 102 reflect the scanner light from their
surfaces, while translucent or transparent documents
102 reflect the light with the aid of the specimen
mount 104.
The reflected light from the specimen 102 is
captured by the camera 110. A charge-coupled device
(CCD) having a linear array of photoresponsive cells
as is well known in the prior art, is included in the
camera 110 to produce analog signals at each pixel
location within a two-dimensional array of pixels
which are proportionate to the pixel reflectivity.
The CCD array within the camera 110 is mechanically
displaced orthogonally to its linear array to effect
a full two dimensiona} scan of the specimen 102. The
camera 110 contains circuitry for effecting a
conversion of tbe analog signals to multi-digital
numbers representing the reflectivity of each pixel
of the document 102. The camera 110 can be a high
speed, full frame TV camera, such as is commercially
available.
-6-
~29~;4:~
Alternate embodiments for the scanning process
are possible. For e~ample, the specimen 102 may be
transparent, such as microfiche, and the specimen
illuminator 108 might, instead of causing most of the
light to be reflected from the specimen 102, shine
light through the specimen 102. The camera 110 would
be located strategically to capture the transmitted
light, and the signals received from the camera 110
would be a measure of the transmissivity of the
specimen 102 rather than the reflectivity.
At the heart of the system 100 is the computer
console 112. In the present embodiment the console
112 contains the central processing unit (CPU),
memory to store the digital images and other data, as
well as program instructions to direct the operation
of the separate components of the digital imaging
system 100.
The user initiates and ultimately controls the
operation of the digital imaging system 100 from the
keyboard 114. Special function keys 116 are provided
for sending special instructions to the system 100
during the digital imaging process.
A monitor 118 is included in the system 100.
The monitor 118 contains components, including a
screen or cathode ray tube (CRT), necessary to
-7-
1296A29
display images from the specimens 102 either directly
from the output of the digitizing process or from
other memory after t~e diyitizing process has taXen
place and been stored.
The present invention includes a document
processor or scanner means 126 ~or automatic scanning
and processing of a stack of documents. The document
processor 126 combines the features of the scanner
106 and some of the functions performed by the
console 112, although the console 112 may control the
overall operation of the digital imaging system 100.
The document processor 126 may also be connected to
another computer system through networking. Many
documents can be more efficiently digitized without
the need for user intervention using the document
processor 126 rather than the scanner 106. Thus, the
present invention anticipates that most of the
digitization will occur using the document processor
126. The scanner 106 is still needed to scan
documents which either cannot fit into the document
processor 126 or are unsuitable for feeding into the
document processor 126.
Figure 2 shows a distributed data-processing
system or distributed processor system 200 which is
functionally equivalent for purposes of the present
.
~ -8-
:: .
~296~29
invention to the system 100 in Figure 1. The core of
the digital imaging file processing system 200 is the
computer 202 which contains the system~s main
processor. The other components of the system 200
are interfaced with the computer 202 and each other
with data link means 204. The data link means
include hardware and software necessary for the
synchronous or asynchronous operation of the
components of the system 200. The distributed
processor system 200 for the present invention
contains a document processor 206, which is identical
to the document processor 126 in Figure 1.
An optical character recognition device or means
208, described infra, is used by the system 200 for
character recognition operations using digital images
from the document processor 206. Main data storage
of the digital images is accomplished using one or
more storage means--here, optical disks 210 and one
or more magnetic disks 212. Various graphic copies
of the stored digital images can be output to a
printer 214, used for printing the digital images~
User int-raction with and initiation of the
system 200 is effected through data linked
workstations 216. The system 200 is capable of
accepting a number of workstations 216, which are
: _9_
1296429
part of the distributed data processing system. In
the preferred embodiment the workstations 216 have
"intelligent" capabilities, and essentially comprise
the following components of the system 100 in Figure
1: a console or central processing unit 112: a user
keyboard 114 and a monitor 120, for the display of
information to the user.
Figure 3 is an alternate embodiment of the
digital imaging file processing system 200 in Figure
2, implemented using a local area network 302, which
connects the major components of the system 200.
Figure 4 shows a diagram of the steps involved
in processing the documents 102 by the system 200.
The digital imaging o a stacX o documents begins by
loading the documents into the document processor
126, shown in mechanical detail in Figure 5. In the
preferred embodiment, a stack of documents to be
digitized will have a header sheet or cover page 600
as shown in Figure 6. A detailed description of
Figure 4 will be given infra.
The sample header page 600 in Figure 6 contains
a bar code 602 at the top to identify the page and
the user. For example, the bar code 602 may
represent fourteen unique digits of a code. The bar
code is read by the~ scanner 126 as explained infra.
. ~: .
10-
.
: ~ :
~ ~ -
1296429
The leftmost two digits may indicate to the document
processor 126 that the par~ic~lar sheet is a header
page so that special information about the stack Oe
documents to be processed can be anticipated.
The next eight leftmost digits of the bar code
602 might represent a unique user identification
number 616. The user identification number 616 may
have many uses. Accordingly, part of the eight-bit
code ~ay include code indicating whether a particular
user is ~o have access to the digital imaging file
processing system 200. The user identification
number 616 is also useful in the classification and
storage of the processed documents. Finally, the bar
code input can be used to track user practices such
as frequency of operation. That type of information
is valuable for billinq and security systems.
The last four digits represent a form number
618, indicating which header form the user is
employing in the digitizing processing. Many
different types of header pages are possible, with
each page indicating different kinds of information
the docu~ent processor 126 is to pay special
attention. A decimal numeric representation 604 of
the bar code 602 is placed below the bar code 602 for
easy identification by the human user.
--11--
i296~29
The sample header page 600 contains four
reference lines--two solid horizontal (606 and 608)
and two broken vertical (610 and 612)--which serve
two purposes. First, by detecting the lines 606-612,
the document processor 126 knows that special header
page information is to follow. Second, the reference
lines 606-612 serve to inform the document processor
126 that special information boxes are located at
known distances from the horizontal and/or vertical
lines. Once the reference linas 606-612 are detected
the document processor 126 scans the information
boxes to find those which are shaded, A shaded box
indicates a particular option chosen by the user.
In the preferred embodiment the scanner control
options are identical for each header page form,
regardless of the user. The first three options
622-626 under scanner control options 620 control the
scanning resolution. For example, the scanninq
resolution may be 200, 300 or 400 dots per inch.
The next two option boxes 628 and 630 indicate
whether the system is to pay special attention to
highlight (yellow) marks in the stack of documents to
be processed. The first of the two boxes 628
indicates that no highlight marks are to be read by
~ the system, wbile the second of the two boxes 630
:
-12-
~2964Z9
indicates that the system is to look for and process
highlight marks. The highlight mark detection means
will be explained infra, with the aid of Figures 7
and 8. The last three boxes in the group 632-634
give information about the optical quality of the
documents in the qroup represented by the particular
page. If the document or documents in the group are
rather faint in nature, the "light original" box 632
is blackened. If the document or documents in the
group are dark in nature, the "dark original" box 636
is blackened. Finally, the "normal original" box 634
is blackened if the documents in the group are of
normal luminosity. These boxes adjust the scanning
circuit and/or algorithms accordingly. The default
choices for the boxes under "SCANNER CONTROL
OPTIONS", numbered as 620, are indicated by an
asterisk. In other words, if the user fails to
specify a box in any of the groups of boxes, the
system 200 will automatically assume that the default
choices are to govern.
The option boxes under the heading "OCR
OPTIONS", numbered as 638, are to be included on all
header page forms as well as the "SCANNER CONTROL
OPTIONS". The boxes under "OCR OPTIONS" control the
actuation and manner of operation of the OCR device.
-13-
i~96429
The first bax, "no OCR" (numbered 640), indicates
that the OCR deYice 208 is not to be used by the
system 200 in processing the particular batch of
documents. The second box, "OCR on highlight areas
only" (numbered 642), indicates to the system 200
that the OCR device is only to be actuated in the
confines of yellow highlight marks. This may
indicate that the user prefers that only the
information appearing as highlighted is to be stored
in the form of alphanumeric characters.
If the user desires the entire digitized
document~s) to be converted to alphanumeric
characters where appearing in the specimen(s) the
"OCR entire document" (numbered 644) is blackened.
The remaining two boxes 646 and 648 under the "OCR
OPTIONS" heading are blackened according to which, if
any portion the user desires the OCR device to
ignore. If the "delete image" box 646 is blackened
the system will only save the text data while
disgarding the non-text data after the document is
completely processed. The "delete image" box may
alternately be defined to mean the OCR will only
convert digital data into alphanumeric characters in
those areas of the specimen which are classified as
primarily textual in nature, while ignoring those
-14-
1296429
70a40-los
area~ of the specimen whlch are clas~lfied as prlmarlly
photographlc.
lf the u~er hlacken~ the "delete text" box 6~8, the system
will treat the resultlng data from the document ln a manner
opposlte to the ~delete lmage" case, above. A method of
dl~tlnquishing between textual and photographic areas ln the
dlgltal lmaging proces~ i~ disclo~ed in Canadian Patent Application
Serial Number 522,846 flled November 13, 1986. The inventlon
dl6closed ln that appllcatlon ha~ been as~igned to Wang
Laboratorles, Inc. of Lowell, Massachu~etts, also the assignee of
the present lnvention.
All of the boxes 652-664 under the headlng ~FILING OPTIONS~
(number 650) are unique to the particular header page form belng
used. The user may create various options under thls category to
be remèmbered by the 3ystem 200. The ~FILING OPTIONS~ determlne
the de~tinatlon and~or clas~ification of the digitized images. The
boxes shown under this headlng ln Flgure 6 are lllustratlve.
Flgure 5 wlll now be descrlbed ln detail. As can ~e seen
fro~ a cooparison of Flgure~ l and 5 the docu~ent processor 126
dlffers ln some detail~ (e.g. physlcal dimensions and number of
trays) from one drawing to the other. However, in all essentlal
respect~, the document processor of Figure S is identical to that
of Flgure l and for that reason the ~ame reference numeral, namely
126, 18 used ln Flgure 5. The document processor 126, powered by a
power supply 500, contalns a document stack 502 for face-down
stacklng of a group of documents to be digitized.
J.'-i96429
The document stack 502 is located at an incline to
the horizontal so that the documents tend to slide
towards the document load lever 504. The document
load lever 504 operates such that only the bottom
sheet of the documents in the document stack 502 is
allowed to pass at one time.
A document feed motor S06 powers a document feed
roller 508, which advances the specimen through an
elbow column 510 for curving the specimen around to
the document alignment mechanism 512. The document
alignment mechanism 512 essentially consists of two
document alignment rollers 514 surrounded by a band
515. The document alignment mechanism 512 is powered
by a drive motor 516 through a belt 517. After the
document has been advanced by the document feed
roller to the document drum 518, the document
alignment mechanism 512 insures that the edges of the
document are flush with the outer surface of the
document drum 518. The document alignment mechanism
512 corrects any misalignments by sending the
document between a document alignment plate 519 and
the band 515. The coefficient of friction between
the document and the band 515 is greater than the
coefficient of friction between the document and the
document alignment plate 519, so that the document
'~
. -16-
~296429
travels with the movement of the band 515 until the
edges of the document reach the document drum 518.
The document drum 518 is rotated by a document
drum belt 520 powered by the document drum drive
motor 522. A special document clamp 524 is located
on the surface of the document drum 518, and is
parallel to the axis of the drum. The document clamp
524 possessès a unique biased clutch feature. That
is, the document clamp only engages when the document
drum is rotated in one direct~on--clockwise in Figure
5. Once the document clamp 524 is engaged, it
continues to be engaged as long as the document drum
518 rotates in the clockwise direction, or stops
rotating. The document clamp 524 is disengaged by a
slight counter-clockwise rotation of the document
drum 518.
Just before the scanning process begins the
document is positioned--with the document feed roller
508 and the the document alignment mechanism
512--inside of the small document clamp 524 until the
document is flush with the back of the clamp 524. In
practice the depth of the clamp may be very small.
In the document receiving mode of the document drum
518 the document clamp 524 is in the disengaged
~ : .
st~te. The document drum 518 then begins rotating
-17_
1296429
clockwise in ~igure S, and after a small angle, the
document clamp 524 engages and clamps the edge of the
document. The clamped document is rotated clockwise
past the document illuminators (lamps) 528 during the
scanning process. The document mainta;ns contact
with the document drum 5~8 by the clamp 524 and
further with the aid of a document press mechanism
526. The document press mechanism may be a simple
ledge of resilient material attached to the document
alignment plate 519, biased toward contact with the
document drum surface. The document press mechanism
insures that the document is properly located on the
surface o~ the document drum 518 during scanning.
During the actual scanning o a document the
document is illuminated by the document illumination
means 528. The resulting image of the document is
reflected by a fold mirror 530 to an objective lens
532 for focusing the line image of the document on
the linear surface of a charge-coupled device (CCD)
array 534. As explained supra, the CCD array
converts each scan line of the image into a line of
analog voltage levels as the drum rotates the
document past the stationary CCD. The image is
digitized using digital circuitry on boards 536 and
538 along with the computer 202 of the system 200.
-18-
~96429
The images are further converted into pixels by any
of thresholding processes well known in the art. The
circuit boards 536 and 538 contain digital hardware
as well as "firmware" to accomplish the
thresholding/digitizing process.
In addition to outputting signals used in the
digitization of the documents, the CCD array 534 a}so
serves as a bar code reader for reading the bar code
602 of the header page 600 in Figure 6. The various
lines comprising the bar code 602 are converted into
a unique character string based on the signals output
by the CCD array 534 while the bar code 602 is being
scanned.
Electronics for the coordination of the various
motors in the document processor 126 are coDtained at
location 540. The document is rotated during the
scanning process until scanning is complete. A
document stamp or solenoid 542 is available to print
various error messages on a header page or document.
The document is rotated until the document cla~p 542
is positioned between an exit roller 544 and an exit
wall or stripper plate 546, both resting against the
document drum 518, at which point the document drum
518 stops rotating momentarily. The document drum
518 is rotated through a small counter-clockwise
-19-
1296429
angle--enough to disengage the document clamp 524 and
release the document.
The coefficient of friction between the document
and the exit roller 544 is greater than the
coefficient of friction between the document and the
document drum 518. Therefore, the exit roller 544,
powered by the drive motor 516 throuqh a belt 517 is
able to advance the document by a counter-clockwise
rotation. With the document free of the clamp, both
the exit roller 544 and the document drum 518 are
rotated until the leading edge of the document
reaches the stripper plate 546 and is forced away
from the document drum 518. The stripper plate 546
is located close enough to the exit roller 544 and
the document clamp 524 in its disengaged state so
that drum must rotate past the exit wall 546 before
the document clamp 542 again closes. By the time the
clamp 524 again closes the leading edge of the
document is separated from the the surface of the
document drum 518 and is following the stripper plate
546. Consequently, the document is not erroneously
re-clamped to the drum 518 by the clamp 524. The
exit roller 544 continues forcing the document
forward until the document finally lands into the
exit tray 548.
-20-
~!96429
The document drum 518 continues its rotation
back to the start position for receiving another
document. It is then rotated through a small
counter-clockwise aDgle to open is document clamp
524. The document processor 126 is now ready to
process another document.
The document processor 126 contains a control
panel 550 for limited user interface. Various
switches 552-560 are included with the document
processor 126. Switch 552 is a job interrupt key for
interrupting a current processing job. Switch 554
indicates that a document has begun to be removed
from the stack and later that it has moved past the
document feed roller 508. Switch 556 indicates that
the document is ready to be clamped and scanned.
Switch 558 indicates a document overflow--that more
documents are in the exit tray 548 than the document
processor 126 can handle. Switch 560 indicates
whether the width of the documents in the document
stack 502 is eight and one-half inches or "A4". In
the preferred embodiment the documents in a batch
ust have the same width.
The method of highlight mark detection shall now
be discussed. One use of highlight mark detection in
the present invention is to locate specially shaped
-21-
~.296429
highlight marXs on a document which indicate the
title of the document. The digitized version of the
document can then be indexed accordinq to title.
Another use of the highlight detection method is to
indicate key words in the document, which the system
200 will use to further classify and index the
digitized document. A third use of highlight marks
and highlight mark detection in the present invention
is to place and have recognized column separators in
a document which the optical character recognition
device (or optical character recognition means) 208
will use as references to maintain the original
column format of the document to be digitized.
Although the particular method of highlight
detection discussed below is for yellow highlight
marXs, the same method can be adapted to detect
highlight marks of other colors. Figures 8 and 9
represent empirical data about the optical
characteristics of yellow highlight marks. From this
data generalizations are formed which lead to the
method of highlight mark detection in the present
invention. The output of the CCD at a single point
or pixel of a document is governed by the following
equation:
-22-
1296429
(A) O = S(k)I(k)R(k)dk
Where O is the output in volts, S is the spectral
sensitivity of the CCD, I is the spectral
illumination and R is the object spectral
reflectance. S, I and R are all functions of
wavelength, k. In order to facilitate color
detection two outputs are needed at a single pixel.
If the the illumination, I, is properly chosen for
the two scans of the pixel under test, the color
characteristics of the pixel can be determined by
comparing the outputs through what will be generally
referred to as an output comparison means (actually
an algorithm).
The specimen is first illuminated with white
light, which in general contains equal amounts of
red, green and blue visible light. Then the specimen
is illuminated with blue light, which contains only a
component of blue visible light and no components of
red and green. A comparison of the two outputs gives
an indication of the color of the pixel area. For a
white pixel, most of the white illuminating light
.:
will be reflected, while in the case of the blue
illuminating light the output will be approximately
one thiFd of the output for white illuminating light
~:
-23-
1296429
since only hlue light is able to be reflected, for
the case when the intensity of the blue light source
is approximately one third of the illumination of the
white light source. In this case the intensity of
the blue light or the gain of the illumination system
should be increased so that the output O for a white
pixel equals that from the white light.
If the pixel under test is yellow, a different
result is obtained. When the pixel is illuminated
with white light nearly all of the white light or
approximately 90 percent is reflected. However, when
the pixel is illuminated with blue light most of the
illuminating light is absorbed by the yellow
background, making the output of the pixel near zero.
The color of the pixel can be determined by a
ratio of the second output (using blue illuminating
light) to the first output (using white illuminating
light). The results of the comparison are shown in
the following equation:
(B) 2/l 1 if the background is
white
O if the background is
yellow
:
~ -24-
~296429
Using equation B all that need be done is divide
the out,puts to obtain a fraction in order to
determine whether the backqround pixel is white or
yellow. However, subtraction is often much easier
for digital computers than division (division
requires more time than subtraction. It is desirable
to use a subtraction method of comparing CCD outputs
to ease the computation burden of the
arithmetic-loqic unit of the document processor. To
approximate eguation B the preferred embodiment of
the present invention subtracts the second output
(using blue illuminating light) from the first output
(using white illuminating light). The results of the
computation are summarized by the following equation:
(C) l ~ 2 0 if the object is white
O if the object is
yellow
Equation C contains results under ideal or near
ideal conditions. This method of approximation is
premised on the fact that the second output is nearly
equal to the first output when the background is
white. Actually, the document processor 126 need
only determine whether the difference is near zero or
-25_
1296429
significantly different from zero to determine if the
background is respectively, white or yellow.
Therefore equation D, below, is used to compute a
difference, B.
1 2 B
If B exceeds a chosen threshold value, then the
background pixel is determined to be yellow. For
example, the threshold value might equal unity. One
of the advantages of this method is that gray or dark
areas neutral in color are not detected as yellow.
The method of yellow highlight marker detection
is actually computed on a cell of pixels at a time
rather than single pixels. For example, cell may be
a block of 16 x 16 pixels on a document containing
400 dots per inch per scan line. An average output
for the entire cell is obtained first using white
illuminating light; then an average output using blue
illuminating light is obtained. As before in
equation D, the second output is subtracted from the
first output to obtain a value B. If B is less than
the threshold value, the cell is determined to be
"white"~ If B is greater than or equal to the
threshold value, the cell is determined to be
~X964Z9
yellow. The cell size is chosen larqe enough so that
if the cell contains all or part of an alphanumeric
character, there will be enough surroundinq pixels to
affect the average output of the cell so that
erroneous white/yellow determinations are avoided.
Referring to Figure 9, a bit map representation
of the yellow marks 902 and their locations within
the bit map is then formed. It is desirable to
convert those areas 902 determined to be highlighted
into rectangles when the bit map representation shows
otherwise; i.e., shape 902 is not a rectangle. It is
contemplated that all highlighted information should
fit into a rectangle varying in dimension. A
rectangle assimilation algorithm (or shape
assimilation means) is applied to the bit map to
obtain rectangles 1002 ~as shown in Figure 10) of the
yellow marks. ~his technique sometimes requires that
some cells previously categorized as un-highlighted
be re-classified as highlighted, while in other
instances a cell classified as highliqhted is
re-classified as un-highlighted.
- After the highlighted cells are assimilated into
rectangles, the rectangles are analyzed for their
geometric shapes with a shape determination means.
The shape determination means is a combination of
-27-
~296429
software designed to search for certain predefined
shapes. For example, if a combination of rectangles
yields an elbow shaped mark 1102 as in Figure 11, it
is deemed to indicate the presence, for example, of
the title of the document. For example, the document
processor 12~ will determine that the title of the
document begins slightly to the right of the vertical
~eg of the "elbow mark", and slightly below the
horizontal leg of the "elbow mark". The title mark
is determined by comparing the original highlight
mark data with the assimilated rectangle. If a
significant portion of the assimilated rectangle was
not included in the original highlighted mark, the
highlight mark is determined to be an "elbow mark".
Another significant rectangle is the horizontal
rectangle 1002 in Figure 10 indicating the presence
of a key word within its boundaries. Still another
significant rectanglc is the long, vertical rectangle
1204 as shown in Figure 12, which indicates to the
OCR device (where OCR operations are to be performed)
that the scan lines of a page 1202 are to be
separated into columns 1204 beginning and ending at
the highlight marks.
Returning now to Figure 4, the process is
started by entering a prospective document to be
-28-
i296429
digitized into the document processor 402: e.g.,
scanners 106 and 126 in Figure 1. Before any of the
documents in a stack are digitized, the document
processor 402 sends the various information from the
header page 600 in Figure 6 to the document processor
server 404, a sotware module. The server 404
includes a bar code reader algorithm for reading the
bar code 602 on the top of the header page 600. The
document processor server 404 seeks verification of
the user identiication number 616 listed on the
header page 600 by sending the identification number
616 to the header page/identification file 406, a
data storage file, If verification is received from
the storage file 406, the document processor server
404 continues the digitization process. If
verification is not received, the documents are
rejected by the document processor 402, and no data
from these documents is sent to the computer 202,
while the respective header page 600 is marked with
the document stamp 542 with "INVALID I.D. ~UMBER" or
similar message. If the user desires to create a new
header page 600 for the system 200 or a new
identification, the header page and identification
number generator 408, a software module, can be made
to generate a new identiication number 616 a take
-29_
~296429
cognizance of the new header page 600 being read.
The document processor server 404 sends image
data to the image access library 410. The image
access library 410 creates a new file for the image
data and sends the corresponding image data each time
a new header page 600 is indicated by the document
processor server 404. The image file 412 contains
not only a bit map representation of the documents,
but the presence or absence of highlight marks, as
well a~ their shapes and locations. Along with
sending the image data to the image access library
410, the document processor server 404 sends the
header page form number 618, the boxes specified on
the header page 600, the user identification number
and the name of the file in the image file 412 where
the corresponding image data has been sent to the
application software module 414. These items have
been stored by the header page/identification number
file 406.
The application software 414 makes a query of
the header page/identification number file 406 as to
what functions to perform. In the present invention
a "personal filing system" message instruction is
received by the application software 414 from the
file 406. This means the filing features of the
~::
-30-
i296429
system 200 explained heretofore are chosen. It is
possible for the application software 414 to perform
other functions unrelated to the present invention.
The application software 414 sends the image
file name, the text file (to be discussed infra) na~e
and the various indexing information, including the
title and ~ey words to the destination software 416.
The destination software is accessed whenever a
document or group of documents is to be available for
editing, observation, etc.
The OCR device, shown in Figure 4 as reference
number 418 is controlled by the OCR device server
software 420, which serves as an interface between
the OCR device 418 and the rest of the system 200.
If OCR operations are chosen, the application
software 414 sends the image file name of the
respective document along with the various OCR
options (listed on the header page 600) to the OCR
device server 420. The OCR device server 420 sends
the image data received from the image access library
:
410 (received from the file name in the image file
412 indicated to the OCR device server 420 by the
application software 414) to the OCR device 41~ for
character recognition of the data. As the image is
: ~ :
being converted by the OCR device 420 text data is
~:: :`
:::
::
-31-
1296429
sent ~ack to the OCR device server 420 until the
conversion process is complete.
The converted data, now in textual form, is
converted by a module 422 into a word processing
document, such as is compatible with the particular
system 200 for word processing. From there the word
processing document is stored in an output or text
file 424. The word processing document can be
accessed by the components of the system 200 via the
computer access package 426.
Other applications of the OCR device 418 are
possible in the system 200 without the use of the
document processor 402 by the image application
module 428.
Various modifications and variations of the
foregoing described invention are obvious to those
skilled in the art. Such modifications and
variations are intended to be within the scope of the
present invention. The embodiments described are
representative of a multitude of variations without
changing the essence of the apparatus operation. For
example, the user may enter some of the same input
information found on a header page, both before and
after digitization, through a workstation 212 in
Figure 2. This includes such information as the
-32-
1296429
location, size and shape of highlight marks used to
identify title and key words, as well as
identification nu~bers, etc.
33-
:: :
