CN1429450A

CN1429450A - Method and system for form recognition and digitized image processing

Info

Publication number: CN1429450A
Application number: CN01803614A
Authority: CN
Inventors: A·P·伯格森; S·范思克; D·格森伯格; A·小列农; S·B·里德
Original assignee: TIMBREL SYSTEMS CORP
Current assignee: TIMBREL SYSTEMS CORP
Priority date: 2000-01-11
Filing date: 2001-01-09
Publication date: 2003-07-09
Also published as: JP2004504650A; CA2397151A1; BR0107575A; WO2001052519A1; AU2001226366A1

Abstract

A method and system for automated identification of pre-printed forms (45), based on the location of a mark (230) made manually on the form (45). The method comprises the steps of receiving digitized ink image data (412), identifying the existence and location of a mark (230) and determining any misalignment between the received ink data and a stored image of the form, and shifting the ink data to correct the misalignment. The system comprises a digitizer (35) and a pre-printed form (45) comprising one or more pre-printed indicators unique to the form indicating where a user is to enter one or more identifying marks (221, 230). The digitizer (35) transmits image data to a computer processor (25), coupled to a database, including one or more digitized images of marks made on a form. The processor (25) identifies the existence and location of one or more identifying marks made by the user, and identifies the form based on the location of said one or more marks.

Description

The method and system that identification of seal page or leaf and digitized image are handled

Invention field

The present invention relates generally to use the data input, particularly identification of a digitlization converter apparatus to be placed on a seal page or leaf on the digitizer and definite system and method that reaches the digital picture skew of correcting digital converter board.

Background of invention

Automatically it is known discerning the system that is placed on the seal page or leaf on the digitizer device.(employed here term " seal page or leaf " relates to a printing image on the space of a whole page (page), rather than an actual space of a whole page; Just, if, then these two pages are regarded as identical seal page or leaf) with two pages of identical image printing.For example, U.S. Pat 5869789 (Reid-Green) discloses a kind of page recognition system, and when being placed on a piece of paper on the digitizer system, this system detects the page number (page number) of encoding on pre-this paper.This system utilizes a built-in scan device to detect a precoding page identifier on this page back side.U.S. Pat 5555101 people such as () Larson discloses an automatic page recognition system, and this system is used for the seal page or leaf that facsimile machine receives, and has used the pattern of preprinting of a seal page or leaf of an identification.

U.S. Pat 5629499 people such as () Flickinger discloses a kind of clipboard, and discloses the methods of used seal page or leaf on two kinds of these clipboards of identification.In first method, a trigger is used to trigger difference seal page identifiers shown on the clipboard; In the second approach, a bar code reader is read the bar code of preprinting on the seal page or leaf.Trigger showing needs clipboard to comprise a microprocessor, and when using one group newly to print page, perhaps this microprocessor needs reprogramming each.

People have known that Digital notebooks writes down the symbol of made on the paper seal face that is placed on it.For example, a kind of CrossPad 35 (referring to Fig. 1) that gone on the market of A.T.Cross company, that is, and a portable Digital notebooks that utilizes wireless transmission to come digitlization and record action of writing.CrossPad sends to one by one people's computer with the action of writing of record and handles and analyze.Yet CrossPad itself can not discern the seal page or leaf of preprinting.Only comprise by the user for the information of computer by the CrossPad recording and sending and on a seal page or leaf, write the stroke of importing.The information that is imprinted in advance on the page or leaf can not be detected by CrossPad.The CrossPad neither one can not have bar code reader by the seal page or leaf ID display of user's triggering yet.

Because only detect information by user writing such as the device of CrossPad, therefore this device can not use in according to the system that preprints information Recognition seal page or leaf usually.When being placed on a plurality of seal pages or leaves on the digitizer, even even the disclosed system of Reid-Green can not detect a seal page or leaf identifying information.For example, Reid-Green will be merely able to discern the seal page or leaf that is positioned at the stack paper bottom.Perhaps, the disclosed device of people such as Flickinger can be read a plurality of pages, but this device need increase (costliness) the seal page reader structure that does not appear on the CrossPad type device.In digitizer, use bar code reader or other seal page reader structure to increase the cost and the weight of digitizer.

Another shortcoming of CrossPad device is that the spatial relationship of writing sign on the paper seal page or leaf is different from the spatial relationship that the respective digital sign is gone up on basis " numeral seal page or leaf ".This deviation (" digital displacement ") can have several millimeters.

Another shortcoming of CrossPad type device results from the following fact: when using a seal page or leaf, the information of preprinting on this seal page is not the part of the digital ink file uploaded.In order to make the digital ink file upload behind efficiency the best, the image of the language ambience information that presents on the paper seal page or leaf need be added on the digital ink image in any time of using.These two images must be aimed at really, so that can rebulid the image of the original seal page or leaf with the information of writing thereon.

Therefore, need a kind of system of automatic seal page or leaf identification of the seal page or leaf that can use on CrossPad and similar device, this system does not rely on the information of preprinting on the seal page or leaf.

In addition, also need a kind of like this system, this system does not need to be integrated into LCD display (connecting same programmable microprocessor) and trigger or other input unit of a heaviness on the CrossPad type electronic writing plate.

In addition, the method and system that also needs the digital displacement of a kind of CrossPad of correction type device and other digitizer.Also need a kind of method that on the seal page or leaf of a image, makes up and aim at the digital ink image of hand-written information with basis seal page or leaf.

Summary of the invention

The present invention includes a kind of method of preprinting the automatic identification of seal page or leaf, comprise following identification step: be identified in and preprint an existence and the position that seal page or leaf is gone up the sign of manual manufacture, and the position of the sign by relatively entering Database Lists seal page or leaf and the relevant position of distinguishing mark are discerned and preprinted seal page.The present invention also comprises a kind of system that preprints the identification of seal page or leaf, comprising: a digitizer; Preprint seal page or leaf for one, comprise one or more designators of preprinting that described seal page or leaf just has that have only, indicate the user to import one or more distinguishing marks; With a computer processor that is connected with database, this computer processor receives one or more digitized images that the seal page or leaf is gone up the sign of making, the existence and the position of one or more distinguishing marks that identification is made by the user, and the described seal page preprinted is discerned in the position of the one or more signs by relatively entering Database Lists seal page or leaf and the relevant position of distinguishing mark.The present invention also comprises a kind of computer software of preprinting the automatic identification of seal page or leaf, comprise: receive and storage is described and preprinted the software that the seal page or leaf is gone up the data of the existence of sign of manual manufacture and position, and the software of described seal page is discerned in the position of the sign by relatively entering Database Lists seal page or leaf and the relevant position of distinguishing mark.

The present invention also comprises the correcting digital displacement and with the method for digital ink File mapping to the numeral seal page or leaf.Usually, a digitlization converter board (for example, CrossPad or similar device) can suffer the upward space parallax between the relative position of respective digital sign of the relative position write on a certain amount of digital displacement-paper seal page or leaf and basis numeral seal page or leaf.Digital displacement amount and direction are constant for each plate (and at given seal page or leaf Anywhere), but have nothing in common with each other to another plate from a plate usually.

The software of preferred embodiment is by determining digital displacement amount and direction with reference to the known boundary mark on the given seal page or leaf.A boundary mark like this is seal page or leaf distinguishing mark preferably in the following discussion.In a preferred embodiment, seal page or leaf indications is a circle.Compare by the center of paper being printed respective digital sign on page known location that goes up seal page or leaf indications circle central authorities and the digital seal page or leaf bottom, determine digital displacement, determine the deviation of level and vertical direction then.In case determine this digital displacement, just it be applied to all figure notations of this seal page or leaf as correction factor.

This correction factor is convenient to the aiming at of image of handwritten numeral ink image and basis seal page or leaf.This aligning is at first realized by discerning the seal page or leaf that writes digital ink data thereon.From database, take out the prestrain image that is identified the seal page or leaf then.This correction factor is used for an ink data and aims at numeral seal page or leaf.Numeral seal page or leaf is aimed at the correction of digital ink when duplicating of paper seal page filled in just as its quilt.Proofread and correct to aim at composograph is handled by optical markers identification (OMR), optical character identification (OCR) and the image segments that deposits database in.

The present invention also comprises a kind of method and system of creating the startup identification of seal page or leaf, is used to create the startup identification of the seal page or leaf of not preprinting seal page or leaf distinguishing mark thereon.In this preferred embodiment on the one hand of the present invention, one has the metal in three millimeters holes or the side that plastic guide device (guide) is secured at the digitizer plate, and wherein said three millimeters holes are positioned on the different choice position of this miter guide.This miter guide preferably is secured on link user's the precalculated position.Miter guide is fixed on the precalculated position of digitizer plate one side and can makes the user in a hole, make sign, wherein, the position that the position in this hole should have corresponding to a pre-franking.Just, this sign has and fills up the same position that sign should have preprinting seal page or leaf identification circle.This allows to preprint not, and the client of the seal page or leaf of printing seal page or leaf distinguishing mark uses the remaining seal page or leaf of above-mentioned openly system.

Brief Description Of Drawings

Fig. 1 has shown each parts of the preferred embodiment of digit preference converter system;

Fig. 2 and Fig. 3 show and use data collection seal page or leaf in a preferred embodiment of the invention;

Fig. 4 is the flow chart that shows the software operation of handling digitalized data;

Fig. 5 and Fig. 6 show the flow chart of the operation that shows seal page or leaf identification software;

Fig. 7 shows the area arrangements that is used for the preferred embodiment of the present invention;

Fig. 8 shows the graphical user interface that is used for the preferred embodiment of the present invention;

Fig. 9 shows the step of the main seal page or leaf registration process of using in the preferred embodiment;

Figure 10 shows the coordinate definition that is used for preferred embodiment;

Figure 11 shows the processing of searching the ink bubbles center of preferred embodiment;

Figure 12 shows the processing of distinguishing alignment image of preferred embodiment;

Figure 13 shows the ink that the merging of preferred embodiment aims at and the processing of reference picture;

Figure 14 shows the seal page or leaf identification of preferred embodiment and handles;

Figure 15 shows the seal page identifiers template of preferred embodiment;

Figure 16 shows the digitizer plate with three side paper guide grooves.

Detailed description of the invention

In a preferred embodiment, as shown in Figure 1, system and method for the present invention is used in combination with a digitizer system, and this system comprises: a digitizer 35, such as CrossPad; Preprint seal page or leaf 45; People's computer 25 and bundled software one by one.For convenience of explanation, used here term " CrossPad " relates to the applicable any digitizer plate of the present invention here.Those skilled in the art will recognize that this system and method can be applicable to any digitizer system, wherein, a seal page or leaf can be placed among the digitizer field.In addition, although relate to the seal page or leaf that prints on the paper in this explanation, those skilled in the art will be appreciated that disclosed system and method is equally applicable to the seal page or leaf of implementing on other media.For example, invention of the present disclosure also works in the seal page or leaf on plastic laminate plate or " Electronic Paper ", such as, by (the Palo Alto of Xerox company, California) Kai Fa " Gyricon " product is perhaps by E Ink (Cambridge, Massachusetts) like products of company's exploitation.

The general operation of digitizer system shown in Figure 1 is as follows.The data collection seal page or leaf of preprinting 45 is placed on the CrossPad35.The user uses CrossPad 35 required handtailor pens that sign is placed on the seal page or leaf 45.CrossPad 35 will print the sign of making on the page or leaf 45 and be recorded as digitized image.The user is connected to CrossPad 35 on the computer 25, and CrossPad35 sends to computer 25 to the image of storage.In an alternate embodiments, but CrossPad35 connects a converter that the Crosspad data conversion is become the Internet compatible format.This converter sends to computer on the remote location via the Internet with the CrossPad data of conversion.In any one embodiment, the Crosspad image of storage is handled by saved software on the computer on the computer 25.Employed seal page or leaf of preferred embodiment and software are described below.

Fig. 2 shows employed first data collection seal of preferred embodiment page or leaf 45.Seal page or leaf left margin is preserved for the identification of seal page or leaf apart from last zone.The designator 230 that " Start Here " indication 225 reminds users to be used to print page identifiers should be filled in, and (circle is a designator; By the sign of filling in the circle making is the seal page identifiers).For each different seal page or leaf, seal page or leaf border is known symbol designator 230 and is arranged in the different local of zone 210.

Used seal page or leaf in the printing preferred embodiment makes them can be used for one or more data entry systems.For example, if with a light scanner scans seal page or leaf image, then seal page or leaf bottom right graticule 215 and seal page or leaf are gone up the image that left graticule 220 is used for aiming at the seal page or leaf.

Fig. 3 shows second data collection seal page or leaf that uses in a preferred embodiment.Seal page or leaf among Fig. 3 does not show label, so that more clearly illustrate the appearance of the seal page or leaf that uses in the preferred embodiment.Need to prove that the seal page identifiers designator bit of seal page or leaf is in the different places in the zone 210 different with seal page identifiers shown in Figure 2 230 among Fig. 3.

Preferably, when the user began to write on a seal page or leaf at every turn, this user filled in a seal page identifiers designator.Yet in order to reduce errors, before CrossPad being transferred to one second user, first user preferably fills in a plurality of designators.For example, first user can use a template that comprises the designator that is used for all seal pages or leaves.If use seal page or leaf 1,3 and 6, first user is placed on template on the template of CrossPad, and fill in the designator that the user prints page or leaf 1, knock the Next Page button on the CrossPad, fill in the designator that is used to print page or leaf 3, knock Next Page designator once more, fill in the designator that is used to print page or leaf 6.First user knocks Back Page button twice then, and CrossPad is passed to second user with folder seal page or leaf 1,3 and 6 thereon.Subsequently, second user can fill in seal page or leaf 1,3 and 6, does not need to fill in suitable seal page identifiers designator (although when second user pursues the page turning of seal page or leaf, still needing to push Next and Back Page button).

Fig. 4 is a flow chart, has shown the operation of the software that uses in a preferred embodiment.When the user of CrossPad 35 fills in seal page identifiers designator 230, be used for the user and will deposit the CrossPad memory together in the image and the position data of printing any out of Memory of writing on the page or leaf.CrossPad is this information stores " ink data ", and it is the data format that IBM has, and but, also can use any suitable form of these data of storage.

In step 410, the user uploads to the ink data of collecting on the personal computer by CrossPad being inserted a serial link cable (though also can use infrared ray, Ethernet or other jockey) that is connected to computer, and pushes a button on the CrossPad.In an alternate embodiments, ink data at first is transmitted to converter, sends the data to computer on the remote location via (preferably) the Internet then.Computer on the remote location can be a personal computer, but server-type computers (for example, an Internet Server) preferably receives data because people expect computer from multiple source.

On step 412, the ink transmission procedure stored on the personal computer receives the ink data that is transmitted, in step 415, with these data as the ink data file storage on the hardware driver of computer.In a preferred embodiment, the IBM InkManager Transfer program that is provided on the Crosspad is carried out this step, and creates the file that is called " notebook ".

After uploading this ink data file, ink transmission procedure emission ink converter program (, being known as TMDInkExec here), step 417.In step 420, TMDInkExec reads the ink data file of uploading, and by the ink conversion program piece that comprises in the data bank of calling the IBM supply, be IBM Electric Ink SDK, the ink conversion of page become non-proprietorial image file (such as, organize 4 TIFF, but also can use other form).The suitable ink conversion program piece of instruction calls that is contained in the file that TMDInkExec comprises according to IBM Electric Ink SDK.In an alternate embodiments, the conversion from notebook to non-ownership picture format on the step 420 can also take place at the remote location computer: for example use Email or file transfer protocol that ink data is passed to far-end computer, far-end computer (or a plurality of computer) poll (for example receives formation then, e-mail box), and execution in step 420 is handled the ink data in batch that is received.

In case completing steps 420.TMDInkExec program 422 is just notified second program, and promptly TMDInkMonitor further handles the TIFF data.In step 424, the in batch page of TMDInkMonitor program in the ink data file optionally distributes or confirms a theme ID (for example, if these pages are seal pages or leaves of filling in for a given patient, then should import this patient's ID).If use a scanner to substitute CrossPad, then can paste a theme ID label on the seal page or leaf, this theme ID label is distinguished the identification number by system assignment.If the user is unallocated theme ID (CrossPad also is this theme of record ID), then the TMDInkMonitor program is distributed a Default Subject ID (preferably increasing progressively 1 at preceding theme ID).

In step 426, will upload ID by the gross and distribute to accordingly tiff image file by the gross, in step 428, the TIFF data are write a database by uploading ID and theme ID by the gross.In step 430, the TIFF data file is copied from database and is handled by an identification engine then, and this identification engine adopts OMR (signal is read, or signal is distinguished) technology that image transitions is become recognition data.A kind of identification engine is a software, and this software provides the affirmation of serving another software module file set, and the program that view data is converted to (identification) computer readable text perhaps is provided.In a preferred embodiment, can use (the Los Gatos of Caere company, California) Caere Developer ' the s Kit 2000 identification engines of selling, but numerous identification softwares (for example can obtain from the market, can be from California, the Mitek Systems in Santiago, Inc., and Minnesota, the NationalComputer System of Ming Niaipusi, Inc. obtain), in a preferred embodiment, those skilled in the art can also use these identification softwares.

In step 432, TMDInkMonitor calls a Form ID and determines program block, and this program block drives corresponding to the seal page or leaf ID of recognition data.Fig. 5 and Fig. 6 show by Form ID and determine the step that program block is carried out.

Seal page or leaf ID determines that program block uses OMR recognizer piece and OMR filling method.The recognizer piece is a software, and it meets the standard of discerning engine and converts dissimilar view data to computer readable text, and the expectation of filling method regulation appears at the type of a view data within the zone.Some filling method is associated with a unique recognizer piece; Other filling method is then supported by more than one recognizer piece.Some example of filling method is: the multi-font machine is printed, OMR, printscript, bar code, dot lattice printing, OCRA and braille.

The zone is an image-region, until full page, comprises the user's interest character.In a preferred embodiment, described zone is a rectangle, but those skilled in the art will recognize that also and can use difform zone.The view data that the zone is covered is disposed respectively and is handled.General area type is figure (nonrecognition), text and such as other zone of bar code or OMR.For the ease of the identification of the view data in the region area, the expected information that will comprise according to each zone is each region allocation ownership.Basic regional ownership is x-y coordinate, size, recognizer piece, filter, language dictionaries and user dictionary.Language dictionaries is to comprise the file that the language vocabulary element adds rule, and described rule is used for generating word from the language vocabulary element.This document can be used for marking non-dictionary word in the identification text, perhaps non-dictionary word is become acceptable word.Some recognizer piece reference language dictionary is made the identification decision to assist them.User dictionary contains the regular expression of word list, character string and definition font, with the content of complementary language dictionary file.One or more user dictionaries can be associated with a specific region.Regular expression in the user dictionary by designated character type, scope or to one or more character positions in the zone effectively each character define the font of permission, to improve accuracy of identification.Regular expression is also called mask or model.Regular expression is used under the situation of the seal page or leaf class of being everlasting, wherein area contents be in certain can the prophesy mode in.

Seal page or leaf identified region template (or, FIZT) be used for from image identification seal page or leaf.Region template is one group of zone or page ownership (for example, corresponding to a seal page or leaf to be identified) of specifying and storing, and is retrieving and is handling such as the media of dish file or data-base recording after being used for.FIZT forms (referring to Fig. 7) by one group of zone that is disposed in the grid graph.Filter can be added to region class or seal page or leaf level, to improve recognition result.Filter is stipulations, thinks that according to institute as far as possible effectively the result eliminates some classification.The example of filter comprises: numeral, capitalization, lowercase and punctuation mark.

Because main body invention has utilized the OMR technology, therefore in another embodiment, non-standard symbol (such as, one by circle around N-be similar to symbol ) be used to refer on the seal page or leaf for information about.This example for information about comprises: clip and paste the zone; The information that needs correction; With reference to another data file.

The zone of seal page or leaf identification is defined as from left to right and from from top to bottom, shown in the shade of Fig. 7 number.Each zone corresponds to the seal page number in a seal page group that belongs to.According to having these zones of enough spatial placement between the zone, so that make the minimizing possibility of wrong seal page or leaf identification.Seal page or leaf identification that the author is marked on mark outside the circle or the deflection during scanning seal page or leaf will make the mistake.Each seal page or leaf in one group is assigned with a zone corresponding to its seal page number, and preprints one in the range of distribution on the seal page or leaf page and do not fill in point (referring to Fig. 2 and Fig. 3).A table (" some mapping ") is mapped to period a seal page number that prints in the page or leaf group.An example table that is used for a mapping is as follows:

Period	Seal page or leaf ID number
Period	Seal page or leaf ID number	????1	?00300001-01
????2	?00300001-02	????1	?00300001-01
????2	?00300001-02	????3	?00300001-03

Storage should " be put mapping ", so that retrieve in the storage medium such as database table or dish file later on.

Referring to Fig. 5, in step 510, seal page or leaf ID determination module is retrieved recognition data from the Mark Detection zone.In step 520, be first to have filled up a little and a search is carried out in the corresponding region, and from zone definitions, determine in step 530.In step 540, from seal page number (for example, in last table, period ' 1 ' correspondence is printed page ID number ' 00300001-01 ') based on retrieval correspondence the some mapping table of period.If on step 545, also determine seal page or leaf ID number,, show the image based on ink data of a page to the user, and, point out the user to import suitable seal page or leaf ID number in step 555 then in step 550.In step 560, the user imports seal page or leaf ID number.Although a preferred embodiment has used number (numeral) identification seal page or leaf, those skilled in the art will recognize that also can use alphanumeric or other symbol.Processing shown in Figure 5 is shown in more detail among Figure 14 in some aspects.

Referring to Fig. 6, after seal page or leaf ID number is determined by system, perhaps by after user's input, execution in step 645.In step 645, (for example, use optical character identification, perhaps ICR is discerned-be also referred to as to printscript, is used for intelligent character recognition according to ID number further recognition data of seal page or leaf; The method for optimizing of identification is shown in Figure 12), in step 650, with write the recognition data of seal page or leaf of image and ID number as one not audit document deposit in the database.Recognition data is organized according to the classification of seal page or leaf, and at random is presented on user's the inspection screen.Data are to show such as a tableau format, so that check.In step 660, each the recognition data element on the screen is checked in customer inspection and affirmation, in the time of suitably, it is proofreaied and correct.In a preferred embodiment, graphical user interface (GUI) 810 (referring to Fig. 8) can make the user watch data according to " context orientation ": for example, if some data is unrecognized, then the user can have the image of the part of a paper seal page or leaf that shows on screen, an image with ink data of this page on the image that overlaps this paper seal page or leaf is so that present the screen picture that a quilt is write the part of paper seal page or leaf to the user.This can make customer inspection and confirm data element, and do not present the seal page or leaf of being write, and can use and watch all images of being write paper seal page or leaf, this can make the operator check and confirm at a remote station (this remote station be from this website location of writing paper seal page or leaf) again, in addition, when data are for example during medical report, the seal page or leaf is gone up a hand-written context directional screen and is watched and can make the user rebuild the author to make the phychology (tone) that this report makes, and patient's individual character.

In step 665, the data of checking are saved in the database.

Fig. 9 shows an overview of preferred seal page or leaf registration process.In step 910, Fig. 5 that execution is associated with text and seal page or leaf shown in Figure 6 identification are handled.In step 915, if seal page or leaf identification step 910 is unsuccessful, then in step 920, seal page or leaf registration process is ended, and this seal page or leaf is noted as " the unknown ".

If seal page or leaf identification step 910 is successful,, carry out a location ink bubbles center processing (referring to Figure 10 and Figure 11 and related text) for the ink image bubble so in step 925.This step cause calculating the ink image bubble coordinate (XlnkCenter, YlnkCenter).

After step 925, and the skew between calculating benchmark bubble and the ink image bubble (XOffset, YOffset): XOffset=XlnkCenter-XMast; YOffset=YlnkCenter-Y.XMast and YMast are represented by X0 and Y0 respectively in Figure 10.

In step 940, carry out ink and aim at: XOffset and YOffset value are applied to ink image, and displacement diagram picture thus.

In step 950, the ink image that identification has been aimed at (referring to Figure 12 and the text that is associated).At last, in step 960, with the ink image of aligning and benchmark image (having discerned the memory image of the seal page or leaf) merging of identification.

Figure 10 shows coordinate and the region deviding that utilizes in the ink bubbles center processing of preferred location, Figure 11 shows treatment step.

Referring to Figure 10, shown coordinate system has the y that increases along downward direction and sits the border.Benchmark region of search 1020 is determined by the seal page or leaf and the seal page or leaf identified region template of identification.Because this seal page or leaf is identified, therefore in a special datum bubble zone 1010, find so that (Xlnk Ylnk) is the ink bubbles circle 1030 in the center of circle.Benchmark bubble zone 1010 be make the benchmark bubble have the center (X0, Y0), this center also is the center of benchmark coating ink regions 1010.Benchmark region of search 1020 is to determine like this: with ink bubbles zone 1010 along each direction (left side and right) distance X expand of horizontal-extending and along distance Y expand of each direction (upper and lower) vertical extent.

Thereby benchmark region of search 1020 is rectangles, and ((X2 Y1) (goes up right corner), (X2, Y2) (following right corner) and (X1, Y2) (following left comer) for X1, Y1) (upper left angle) to have angular coordinate.

Location ink center processing is displayed among Figure 11.In step 1105, initializing variable.BubbleZoneLeft is the x coordinate in regional 1010 left sides of bubble.BubbleZoneRight is the x coordinate on regional 1010 right sides of bubble.BubbleZoneTop is the y coordinate of regional 1010 top sides of bubble.BubbleZoneBottom is the y coordinate of regional 1010 bottom sides of bubble.Variable Y Scan is set to Y1.

Variable Y Top, YBottom, XLeft and XRight are set to Y2, Y1, X2 and X1 respectively at first.In the ending of location ink bubbles center processing, the y that variable Y Bottom will contain ink bubbles 1030 sits target value, and the x that XLeft will contain the left side of ink bubbles 1030 sits target value, and the y that XRight will contain the right side of ink bubbles 1030 sits target value.It should be noted that this processing is very thick, even ink bubbles 1030 is not a disk completely, this processing also will be determined " the " center " of ink bubbles 1030.

In step 1110, variable X Scan is set to X1.In step 1115, carry out a check, (XScan YScan) (is positioned at (X1, Y1)) at first to check a black picture element whether to be positioned at the upper left angle of benchmark region of search 1020.If if answer is negated then to increase progressively variable X Scan in step 1160.If the value of check on step 1165, then this increases progressively and can not cause a value greater than X2 (that is, if this scanning does not reach the right side of benchmark region of search 1020, then repeating step 1115).If on step 1165, the increasing progressively of the variable X Scan on the step 1160 causes XScan to have value greater than X2, increases progressively variable Y Scan so on step 1170.If the value of YScan so, is calculated XlnkCenter and YlnkCenter greater than Y2 on step 1180 during the value of check on step 1175.If the value of YScan is not more than Y2 on step 1175, then repeating step 1110.Software from left to right and from top to bottom scans benchmark constriction zone 1020.

If answer is yes on the step 1115, then on step 1120, carry out a check, whether the currency of test variable YScan is less than the currency of YTop.If then YTop is set to equal YScan on step 1125.Thereby find the top of ink bubbles 1030, and its y coordinate is defined as the currency of YScan.If the answer is in the negative on the step 1120, then execution in step 1130, if the answer on the step 1120 is sure, then execution in step 1125.

On step 1130, carry out a check, so that the currency of check XScan is whether less than the currency of XLeft.If answer is yes, then XRight is set to equal XScan on step 1150.If answer is sure on the step 1130, then execution in step 1145, if the answer on the step 1130 is negated that then execution in step 1145.

On step 1145, carry out a check, whether so that check XScan greater than XRight, if answer is sure, then XRight is set to equal XScan on step 1150.If answer is negated on the step 1145, then execution in step 1155, if the answer on the step 1145 is sure, then execution in step 1150.On step 1155, YBottom is set to equal YScan.As mentioned above, step 1160 is followed step 1155.When processing was proceeded, the value of YTop was successively decreased, and had reflected that until it the y of the top pixel in the ink bubbles 1030 sits target value; The value of YBottom increases progressively, and has reflected in the ink bubbles 1030 that until it the y of pixel sits target value bottom; The value of YLeft is successively decreased, and has reflected that until it the x of high order end pixel in the ink bubbles 1030 sits target value; The value of YRight increases progressively, and has reflected that until it the x of low order end pixel in the ink bubbles 1030 sits target value; Like this, when last arrival step 1180, the center (if ink bubbles 1030 forms a desirable disk) of ink bubbles 1030 will be accurately determined in the calculating of XlnkCenter (=(XLeft+XRight)/2) and YlnkCenter (=(YTop+YBottom)/2), and the center of approximate (estimation) ink bubbles 1030 suitably under any circumstance.The method at this definite ink bubbles 1030 centers only is a preferred embodiment.Can not be used for limiting the scope of claims of the present invention.Those skilled in the art will recognize that the method that has various known definite erose two-dimensional object centers, under the condition that does not deviate from the open scope of the present invention, and can be with the alternative said method of these methods.

Return Fig. 9,, use formula XOffset=XlnkCenter-XMast and YOffset=YlnkCenter-YMast to calculate ink bubbles 1030 and (have the deviation between the center (XMast, YMast)) with the benchmark bubble in step 930.Then applicable value XOffset and YOffset the displacement ink image.For example, if XOffset for YOffset just for negative, then with ink image to distance of left dislocation | XOffset|, and to distance of bottom offset | YOffset|.

Preferably according to the number of pixels measuring distance.Like this, in a preferred embodiment, use | the mobile ink image pixel in XOffset| pixel level ground, and use | the vertically mobile ink image pixel of YOffset| pixel.Following C++ sign indicating number shows the method for optimizing of carrying out the ink image displacement:

// // // // // // // // // // // // // // // // // // // // // // // // // // // // function: calculate the bubble skew // // calculate the algorithm // parameter of automatic alignment offset: if // in//hllmg BITMAPHANDLE* have the ink pixel // iBubbleTop bubble reference area top that comprises bubble Y coordinate // iBubbleLeft bubble reference area left part X coordinate // iBubbleBottom bubble reference area bottom Y coordinate // iBubbleRight bubble reference area right part the X coordinate // // out//* piOffsetX is set to the XOffset of calculating or is set to the YOffset of calculating or as 0/ of default value/ find bubble as 0//* piOffsetY of default value, if then turn back to 0//discovery bubble, then turn back to 1//int CalculateBubbleOffset (BITMAPHANDLE * hllmg, int iBubbleTop, int iBubbleLeft, int iBubbleBottom, int iBubbleRight, int*piOffsetX,＜!--SIPO<dP n="13">--<dp n="d13"/>Int*piOffsetY) maximum of the benchmark region of search # definition ALIGN_THRESHOLD 50//pixel of the benchmark region of search # definition EXPAND_X 40//X direction of # definition EXPAND_Y 50//Y direction allows to aim at automatically int iRetuen=0: // be not offset piOffsetX=*piOffsetY=0 at first; COLORREF cr; Int x, y; Int iMinY=-1; Int iMinX=-1; Int iMaxY=-1; Int iMaxX=-1; Int iStartY=__max (iBubbleTop-EXPAND_y, 0); Int iStartX=__max (iBubbleLeft-EXPAND_x, 0); Int iEndY=iBubbleBottom+EXPAND_y; Int iEndX=iBubbleRight+EXPAND_x; Bool bFound=mistake; For (y=iStartY; Y＜=iEndy; Y++) { h=0; For (x=iStartX; X＜=iEndX; X++) { // inspection black picture element cr=GetPixelColor (hLlmg, y, x); (if cr==RGB (0,0,0)) {＜!--SIPO<dP n="14">--<dp n="d14"/>If (! BFound) bFound=is correct; // if all coordinates are initialised to (iMinY==-1) { iMinY=y of first black picture element; If (iMaxY==-1) { iMaxY=y } if (iMinX==-1) { iMinX=X; If { (iMaxX==-1) { iMaxX=X; If { // upgrade mins and maxes be the (iMaxX=x of x＞iMaxX); If (the iMaXY=y of x＞iMaxY); If (the iMinX=x of x＜iMinX); If (bFound) { iReturn=1; The center of // calculating ink bubbles (Xlnk, Ylnk)＜!--SIPO<dP n="15">--<dp n="d15"/>Int iBubbleCenterX=iMinX+ (iMaxX-iMinX); Int iBubbleCenterY=iMinY+ (iMaxY-iMinY); Center (X0, Y0) the int iReferenceCenterX=iBubbleLeft+ (iBubbleRight-iBubbleLeft) of // calculating benchmark bubble; Int iReferenceCenterY=iBubbleTop+ (iBubbleBottom-iBubbleTop); // calculate and print page or leaf alignment offset (XOffset, YOffset) * piOffsetX=iBubbleCenterX-iReferenceCenterX automatically; * piOffsetY=iBubbleCenterY-iReferenceCenterY; // make true skew be not more than max piOffsetX=__min (* plOffsetX, ALIGN_THRESHOLD); PiOffsetY=__min (* plOffsetY; ALIGN_THRESHOLD); Other TRACE0 (_ T (" CalculateBubbleOffset no ink found in bubble zone. n ")); Return iReturn; } // ending function: CalculateBubbleOffset/ // // // // // // // // // // // // // // // // // // // // // // // // // // // // //

By merging white page " sky " the image shift ink pixel that loads ink and planned foundation.The computation system by turn that the pixel result is set to 255 maximums (white) is used to merge pixel.Following code is carried out this operation:

CombineBitmapPixels (﹠amp; BmpWhitePage, " white page " pixel data structure (also causing the destination) of // foundation

0, the X source point of // white page pixel

0, the Y source point of // white pixel

BITMAPWIDTH (﹠amp; BmpWhitePage), // whole X pixels in the white page

BITMAPHEIGHT (﹠amp; BmpWhitePage), // whole Y pixels in the white page

﹠amp; Bmplnk, the ink pixel data structure of // loading

IXOffset, the X source point of // ink pixel (X skew)

IYOffset, the Y source point of // ink pixel (Y skew)

CB_OP_ADD); // merge algorithm type

In case ink image is by displacement (aligning), OMR and OCR handle and just carry out (generations)-discern in step 950 image of having aimed at.

Figure 12 shows the preferred inks image recognition processing.In step 1210, load the Field Definition file that is used for a seal page or leaf.In step 1220, ink image is loaded in the memory.In step 1230, Field Definition file and image are sent to the identification engine that is used to discern, in step 1240, the result who discerns is saved on the database.

Turn back to Fig. 9, aim at the ink image of having aimed at that is identified with the benchmark image of (black) seal page or leaf of identification in step 960.Figure 13 shows this processing.The benchmark image that will print page or leaf in step 1310 is loaded in the memory.In step 1320, ink image is loaded in the memory.

In step 1330, ink image and benchmark image are merged.By use the algorithm of OR by turn that merges pixel plannedly, ink pixel and benchmark image pixel are merged.Following code is carried out this operation:

CombineBitmapPixels (﹠amp; Bmplnk, the ink pixel structure (also causing the destination) of // loading

0, the X source point of // ink pixel

0, the Y source point of // ink pixel

BITMAPWIDTH (﹠amp; Bmplnk), // whole X pixels in the ink

BITMAPHEIGHT (﹠amp; Bmplnk), // whole Y pixels in the ink

﹠amp; BmpReference, the benchmark pixel data structure of // loading

0, the X source point of // benchmark pixel

0, the Y source point of // ink pixel

CB_OP_OR); // merge algorithm type

At last, in step 1340, the image that merges is saved on a dish or other storage medium.

Figure 14 shows the seal page or leaf identification of preferred embodiment and handles.Some processing of this processing also is displayed among Fig. 5.In step 1410, a seal page or leaf is collected Form ID zone definitions file load in memory.In step 1420, an ink image is loaded in the memory.On step 1425, call an identification and use zone definitions and the ink image that loads.In step 1430, variable FormID adjusted to equal 1.In step 1435, carry out a check, so that the character in the recognition result on the check FormID equals 1.If equal 1, then this seal page or leaf is identified, and finishes this processing in step 1450.If be not equal to 1, execution in step 1440 so, increase progressively FormID in this step.In step 1445, carry out a check, so that whether inspection is greater than a quantity of printing the concentrated seal page or leaf of page or leaf.If, then because nonrecognition seal page or leaf and on step 1455 end process.If not, then handle and turn back to step 1435.

In another embodiment of the present invention, use not printing thereon to print preprinting of page or leaf designator and print page or leaf.In order to use this seal page or leaf, a template (preferably metal or plastic formwork) is pasted on the digitizer plate.This template is as a printing stencil (referring to Figure 15): the user fills in the hole in miter guide; Each hole is corresponding to different seal pages or leaves.In this manner, the seal page or leaf that is applied to not preprint seal page or leaf designator is handled in above-mentioned seal page or leaf identification.

In a preferred embodiment, use all configurations as shown in figure 15.Template is a plastic covered plate 1510 preferably, this overlay is enclosed within (referring to Figure 16, the digitizer plate has three sidepiece paper miter guides 1610 in its left side) on the paper miter guide of the sidepiece on the digitizer plate (preferably CrossPad or its derivative).Template 1510 preferably has three grooves 1530, and these three grooves can make template be enclosed within on three protruding paper miter guides on the digitizer plate.Those skilled in the art will recognize that the digitizer plate that should substitute other configuration with other adherence method.For example, if only there is the paper guiding of a projection on the digitizer plate, and be positioned at the right-hand side of plate, then template 1510 should have a groove that is provided with along its right side.

Turn back to preferred embodiment shown in Figure 15, three grooves 1530 be separated by and be approximately (± 2mm) 48mm, wide is 1mm, apart from the left hand edge 5mm of template 1510, length is about (± 1mm) 48mm.The top of top groove 1530 preferably with the top edge of template 1510 at a distance of 15mm, the bottom of lowermost end groove 1530 is preferably and the feather edge of template 1510 15mm apart.

The template 1310 thick 2mm that are about are about and are 273mm, the wide 27mm that is about.The diameter of each seal page or leaf designator bubble 1520 is preferably 2.5mm, and the preferred distance of the minimum between the bubble 1520 is 4mm.

Although shown in and described embodiment can fully realize theme of the present invention, be to be understood that these embodiment of demonstration are only used for illustrational purpose, rather than the purpose that is used to limit.For example, can be easily CrossPad be revised as making user's page (for example, directly jumping to the page 3) that can jump, moves on the page or leaf and a page or leaf shift button down to substitute repeatedly to knock from the page 14.

Claims

1, a kind of method of preprinting the automatic identification of seal page or leaf may further comprise the steps:

(a) receive and store one of description and preprinting the sign that prints manual manufacture on the page or leaf

The data of position;

(b) according to the described seal page or leaf of the location recognition of described sign.

2, method according to claim 1 wherein receives in step (a) and the data of storage are digital image datas, and this view data is received and is stored in by computer in the computer storage media.

3, method according to claim 2, wherein step (b) comprising: determine a seal page or leaf identified region that comprises the described position of described sign, and determine the seal page or leaf corresponding to seal page or leaf identified region.

4, method according to claim 3, wherein step (b) comprises the table of storage coupling seal page or leaf identified region and seal page or leaf ID.

5, method according to claim 4, wherein step (b) also comprises:

The digital image data that is received is formed one again will be by the form of an identification engine processing;

By handle the described view data that forms again with described identification engine, convert the digital image data that forms again to recognition data;

For one or more seal page or leaf identified regions are retrieved recognition data;

Recognition data for the described retrieval of range searching that comprises sign; With

Discern the seal page or leaf by mating described zone with seal page or leaf ID.

6, method according to claim 5 also is included on the computer monitor step that shows digital image data, makes the seal page or leaf that the user can artificially identification written image.

7, method according to claim 6 also is included on the subscriber computer monitor so that the form of checking shows the step of recognition data.

8, method according to claim 7 also comprises the step of the digital image data on the image that shows the seal page or leaf that is superimposed upon identification, so that manifest the screen picture of seal page or leaf and write on the image that prints on the page or leaf.

9, method according to claim 8 also comprises by described recognition data is compared with the digital image data on the image that is superimposed upon identification seal page or leaf that is shown, and checks the step of described recognition data.

10, a kind of system that preprints the identification of seal page or leaf comprises:

A) digitizer;

B) preprint the seal page or leaf for one, comprise one or more fingers of preprinting that described seal page or leaf just has

Show symbol, with one or more identification markings of indication user input; With

C) computer processor that connects database is used for

I) be received in one or more digitized images of having made sign on the seal page or leaf,

Ii) discern the existence or the position of one or more distinguishing marks of making by the user

Put and

Iii) print page or leaf according to the location recognition of described one or more signs.

11, preprint the computer software of the automatic identification of seal page or leaf, comprising:

A) receive and storage is described one in the position of preprinting the sign of manual manufacture on the seal page or leaf

The software of the data of putting;

B) according to the described software that prints page or leaf of the location recognition of described sign.

12, software according to claim 11 also comprises the software of storing a coupling seal page or leaf identified region and printing the table of page or leaf ID.

13, software according to claim 11, the data with storage that wherein received are to be stored in a digital image data in the computer storage media.

14, software according to claim 13 comprises that also management is used for the software of the described digital image data of follow-up benchmark.

15, software according to claim 14, wherein identification software comprises: determine the software of the seal page or leaf identified region of a described position that comprises described sign, and determine the software corresponding to the seal page or leaf of seal page or leaf identified region.

16, software according to claim 15, wherein identification software also comprises such software:

By handle the described view data that forms again with described identification engine, convert the described digital image data that forms again to recognition data;

By coupling the seal page or leaf is discerned with seal page or leaf ID in described zone.

17, software according to claim 16 also is included on the computer monitor software that shows digital image data, makes the seal page or leaf that the user can artificially identification written image.

18, software according to claim 17 also is included on the subscriber computer monitor so that the form of checking shows the software of recognition data.

19, software according to claim 18 also comprises showing the software that is superimposed upon the digital image data on the image discerning the seal page or leaf, so that manifest the screen picture and the image that writes on the seal page or leaf of seal page or leaf.

20, software according to claim 19 also comprises by described recognition data is compared with the shown digital image data that is superimposed upon on the image of discerning the seal page or leaf, checks the software of described recognition data.

21, software according to claim 20 also comprises according to writing on non-standard symbol on the seal page or leaf and distinguishes for information about software.

22, a kind of method of handling digital image data may further comprise the steps:

(a) receive digital image data;

(b) determine first included in the described digital image data position that indicates

Put;

(c) determine the distance and the direction that indicate the reference position from described first;

(d) displacement of the described digital image data of execution, wherein said displacement is with institute

State distance and direction and be the basis.

23, method according to claim 22, wherein said digital image data is corresponding to the sign that is seated in made on the seal page or leaf of digitizer on board-like.

24, method according to claim 23, the wherein described seal page or leaf of the location recognition in a zone by making a distinguishing mark.

25, method according to claim 24, wherein said first sign is described distinguishing mark.

26, method according to claim 25, wherein said distinguishing mark is approximately circle.

27, method according to claim 25, the described position of wherein said first sign is determined by the center of locating described first sign.

28, method according to claim 26, the described position of wherein said first sign is determined by the center of locating described first sign.

29, a kind of method of handling digital image data may further comprise the steps:

A) receive digital image data;

B) determine first included in the described digital image data position that indicates;

C) determine the distance and the direction that indicate the reference position from described first;

D) displacement of the described digital image data of execution, wherein said displacement is with described

Distance and direction are the basis;

E) by handling described digital image data, with described with an identification engine

The displacement digital image data converts the recognition image data to; With

F) merge described recognition image data and benchmark image.

30, method according to claim 29, wherein said digital image data is corresponding to the sign of made on the seal page or leaf that is seated on the digitizer plate.

31, method according to claim 30, wherein said benchmark image are images of described seal page or leaf.

32, method according to claim 31, the wherein described seal page or leaf of the location recognition in a zone by making a distinguishing mark.

33, method according to claim 32, wherein said first sign is described distinguishing mark.

34, method according to claim 33, wherein said distinguishing mark is approximately circle.

35, method according to claim 33, the described position of wherein said first sign is determined by the center of locating described first sign.

36, method according to claim 34, the described position of wherein said first sign is determined by the center of locating described first sign.

37, a kind of method of preprinting the automatic identification of seal page or leaf may further comprise the steps:

The data of putting, wherein said sign is by pasting on the digitizer plate

A hole of template in fill in and make described sign; With

B) according to the described seal page or leaf of the location recognition of described sign.

38, according to the described method of claim 37, the data that wherein receive in step (a) and store are digital image datas, and this view data is received and is stored in by computer in the computer storage media.

39, according to the described method of claim 38, wherein step (b) comprising: determine a seal page or leaf identified region that comprises the described position of described sign, and determine the seal page or leaf corresponding to seal page or leaf identified region.

40, according to the described method of claim 39, wherein step (b) comprises the table of storage coupling seal page or leaf identified region and seal page or leaf ID.

41, according to the described method of claim 40, wherein step (b) also comprises:

42, according to the described method of claim 41, also be included on the computer monitor step that shows digital image data, make the seal page or leaf that the user can artificially identification written image.

43,, also be included on the subscriber computer monitor so that the form of checking shows the step of recognition data according to the described method of claim 42.

44,, also comprise the step of the digital image data on the image that shows the seal page or leaf that is superimposed upon identification, so that manifest the screen picture of seal page or leaf and write on the image that prints on the page or leaf according to the described method of claim 43.

45, according to the described method of claim 44, also comprise by described recognition data is compared with the shown digital image data that is superimposed upon on the image of discerning the seal page or leaf, check the step of described recognition data.

46, a kind of system that preprints the identification of seal page or leaf comprises:

A) digitizer;

B) preprint the seal page or leaf for one,

C) template that sticks on the described digitizer, described template comprises

One or more holes, the user is by one of the described seal page or leaf of described hole input identification

Individual or a plurality of identification marks; With

D) computer processor that connects database is used for

Put and

47, preprint the computer software of the automatic identification of seal page or leaf, comprising:

The software of the data of putting, described sign be by one of a template or

Fill in making in a plurality of holes; With

C) according to the described software that prints page or leaf of the location recognition of described sign.

48,, also comprise the software of the table of storage coupling seal page or leaf identified region and seal page or leaf ID according to the described software of claim 47.

49, according to the described software of claim 47, the data with storage that wherein received are to be stored in a digital image data in the computer storage media.

50,, comprise that also management is used for the software of the described digital image data of follow-up benchmark according to the described software of claim 49.

51, according to the described software of claim 50, wherein identification software comprises: determine the software of the seal page or leaf identified region of a described position that comprises described sign, and determine the software corresponding to the seal page or leaf of seal page or leaf identified region.

52, according to the described software of claim 51, wherein identification software also comprises such software:

The digital image data that is received is formed a form of being handled by an identification engine again;

53, according to the described software of claim 52, also be included on the computer monitor software that shows digital image data, make the seal page or leaf that the user can artificially identification written image.

54,, also be included on the subscriber computer monitor so that the form of checking shows the software of recognition data according to the described software of claim 53.

55,, also comprise showing the software that is superimposed upon the digital image data on the image discerning the seal page or leaf, so that manifest the screen picture and the image that writes on the seal page or leaf of seal page or leaf according to the described software of claim 54.

56, according to the described software of claim 55, also comprise by described recognition data is compared with the shown digital image data that is superimposed upon on the image of discerning the seal page or leaf, check the software of described recognition data.

57,, also comprise according to writing on non-standard symbol on the seal page or leaf and distinguish for information about software according to the described software of claim 56.