Foothill-Deanzacommunitycollegedistrict

AdvancedOCR

HighTechCenterTrainingUnit

oftheCaliforniaCommunityCollegesatthe

Foothill-DeAnzaCommunityCollegeDistrict

21050McClellanRoad

Cupertino,CA95014

(408)996-4636

(800)411-8954

URLtoourCClicense:

CreativeCommonswebsite:

AlthoughyoucanscanwitheitherOmniPageorFineReader,werecommendthatyouscanyourfilestoTIFF, using the scanning utility that comes with your scanner,andthenworkwiththeresultingmultipageimage.Thereareanumberofreasons:itpreservestheTIFFfilesforlaterusewithotherapplications,itpreventsproblemswithcrashinginthemiddleofscans,itallowsyoutotakefulladvantageoftheoptionsthatare built into yourscanner. Please note that you can combine multiple scanned files (TIFF and JPEG, etc.) within the OCR program.

AbbyyFineReader

ABBYYUSA

47221FremontBoulevard

Fremont,CA94538

510-226-6717

Interface

ToolbarSet-up

GotoViewToolbars.TurnontheQuickAccessToolbar.

Go to View > Toolbars > Customize Toolbars…. Set the Categories to “Image,” and find the Eraser tool under Commands. Click the double-headed arrow to move the eraser tool onto your Quick Access Toolbar.

Choose View > Customize Toolbars. Move the Order Areas icon onto the Quick Access Toolbar.

OptionsSet-up

Inthemenus,gotoToolsOptionsorusethekeyboardshortcutCTRL+Shift+O.

DocumentTab

UndertheDocumenttab,youcanchoosethelanguagesthatareinyourdocument.ClicktheEditLanguagesbuttontoaccessmathematicalandscientificoptions.

Themorepreciseyoucanbeaboutthelanguagesinadocument,thebetteryourrecognitionwillbe.Remembertoinclude“Greek”wheneveryouprocessmathsothatyourecognizetheGreekletters(likepi)thatareusedasmathematicalsymbols.

1. Scan/OpenTab

Makesurethat“Donotreadandanalyze”isunchecked.YouwanttobeabletocontrolwhatFineReaderdoes,nothaveitstartautomatically. Do enable the image preprocessing. That feature allows FineReader to deskew and enhance the scanned image (or PDF). If you have two pages per image page, you can choose “Split dual pages.” It is better to rotate the pages yourself, rather than asking FineReader to do so.

2. ReadTab

UndertheReadtab,youcanchoosethe reading mode (Thorough) and opt to train the patterns of unusual fonts.

FineReadernowbehavesmuchmorelikeOmniPagewithitsabilitytotrainrecognitionofunusualfonts.Inordertousethisfeature,makesurethat“Trainuserpattern”isselected.Toloadanalreadyestablishedpattern,select“Useuserpattern.” Below are the instructions from the FineReader help menu.

SelectToolsOptions…andclickthe2.Readtab.
SelectTrainuserpatternunderTraining.
ClickthePatternEditor…button.
InthePatternEditordialogbox,clickNew….
IntheCreatePatterndialogbox,typeinanameforyournewpatternandclickOK.
ClosethePatternEditorandtheOptionsdialogboxesbyclickingtheClosebuttonineach.
OnthetoolbaratthetopoftheImagewindow,clickRead.
NowifABBYYFineReaderencountersanunknowncharacter,aPatternTrainingdialogboxwilldisplaytheunknowncharacter.
Teachnewcharactersandligatures.

Aligatureisacombinationoftwoorthree"glued"characters(forexample,fi,fl,ffi,etc.).Thesecharactersaredifficulttoseparatebecausetheyare"glued"duringprinting.Infact,betterresultscanbeobtainedbytreatingthemassinglecompoundcharacters.

Note:Wordsprintedinboldoritalictypeinyourtextorwordsinsuperscript/subscriptmayberetainedintherecognizedtextbyselectingthecorrespondingoptionsunderEffects.

Togobacktoapreviouslytrainedcharacter,clicktheBackbutton.Theframewilljumptoitspreviouslocationandthelatesttrained"characterimage-keyboardcharacter"correspondencewillberemovedfromthepattern.TheBackbuttonwillfunctiononlywithinoneword.

Important!

Youcanonlytrainthesystemtoreadthecharactersincludedinthealphabetoftherecognitionlanguage.TotrainABBYYFineReadertoreadcharactersthatcannotbeenteredfromthekeyboard,useacombinationoftwocharacterstodenotethesenon-existentcharactersorcopythedesiredcharacterfromtheInsertCharacterdialogbox(clicktoopenthedialogbox).
Eachpatternmaycontainupto1,000newcharacters.However,avoidcreatingtoomanyligatures,asthismayadverselyaffectOCRquality.

3. SaveTab

UndertheSaveTab,youcanaccesssettingsforparticularformats.

YoucansettheoptionsforhowFineReaderwillsavedocumentsintodifferentformats.ForMicrosoftWord,youcanretainthepagebreaks, if you wish. Please note that it is generally best to use Formatted text for the layout, rather than Exact Copy or Editable Copy, which may result in text boxes.

ViewTab

UndertheViewtab,you can now choose whether to see thumbnails or details in the document window. For those using screen readers, I would recommend using the Details panel.

Both views can be adjusted to include exactly the information that you want to show. Click the Customize button.

Note that in the combo box at the bottom of the page, you can choose which colors to use for highlighting uncertain characters, non-dictionary words, etc.

Advanced Tab

On the Advanced tab, you usually do not want to open the last FineReader document at startup.

SpellChecker Settings

UndertheSpell Checker Settings button,youcanuncheckthe"Stopatwordsnotfoundindictionary"optionifyouwillbespellcheckinganywaylaterinWord.IfyouwillnotbetakingthedocumentintoWord,thengoaheadandturnonthedictionary. Note that if FineReader is not quite getting the spacing around punctuation correct, you can adjust the setting in this window.

You can also choose specialized dictionaries in this tab by selecting the View Dictionaries button. Note that if you are running OCR on a science text related to medicine, biology, or chemistry, it is good to turn on the English Medical Terminology. Similarly, if you are working on a business or legal text, you may receive better results by loading the English Legal Terminology.

ProcessinganImage(TIFForPDF)File

StepOne:OpenanImage File or a PDF File

Click on the Open icon or use File > Open or use the keyboard shortcut Ctrl + O.

StepTwo:AnalyzeLayout

UsetheAnalyzeLayoutoption(Document > Analyze Layout or Ctrl + Shift + E)firstoncomplexlayoutsthenreadthepages.This workflow reduces the chance of the program crashing and allows layout adjustment without running OCR more than once..

StepThree:AdjustAreas

Usethetoolstoaddtoordeleteareas (formerly called blocks).Toreorderzones,rightclickandturn ontheArea Properties. It will display at the bottom of the image window. To change the reading order, simply change the number after the # sign.

Tip: To see more of the page, turn off the Area Properties by right-clicking on the image and deselecting “Area Properties.”

StepFour:ReadDocument

Go to Document > Read Document or use the keyboard shortcut Ctrl + Shift + R.

StepFive:CheckSpelling

Go to Tools > Check Spelling. If you wish to change the settings for the spell checker, you can click on “Options.”

StepSix:SavetheDocument

File > Save As > Microsoft Word Document

Notethatyoucanadjustyoursettingsforaspecificformatbyclickingon“Options.”Toavoidtextboxes,savethelayoutas“Formatted Text.”

FineReaderTips

Zoomwindow:FineReaderhasazoomwindow(ViewZoomWindow)thatallowsyoutoenlargeselectedareasoftheimageortext.

Stopspellcheck:TotelltheprogramjusttofindOCRerrorsandnotunknownwords,gotoToolsOptionsCheckSpelling.Uncheck"Stopatwordsnotfoundindictionary.".

Document, Page, or Area:Controls are now broken out under menu choices based on the size of the overall section that you want to work on. To apply a command to the entire document, look under Document on the menu bar. When you wish to work on an individual page, look under Page, etc.

Eraser:TheerasertoolallowsyoutoedittheunderlyingTIFFfilebydeletingpixels(i.e.,changingblacktowhite).Ifyouwishtosavethechanges,gounderFiletoSaveImages.

The eraser tool is now located under Page > Edit Page Image or use the shortcut keys Ctrl + Shift + C. The image for the page you are on will open in a new window titled Edit Image.

Automating Tasks

You can now set up workflows in Abbyy FineReader.

Creating an Automated Task

As an example, we will create a workflow to OCR TIFF files that are in a folder and then save the files as a FineReader file. All files will be saved in one FineReader batch.

Start the Automation Manager (select Tools>Automation Manager or Ctrl + T).

In the Automation Manager dialog box, click the New button on the toolbar.

In the Automation Wizard dialog box, click the desired step from the list of steps in the left-hand pane. The selected step will be displayed in the right-hand pane.

Click Change… to change the properties of the step. In this case, click Change and then Browse to find the folder containing the files you wish to process.

If you want to be able to make adjustments to the layout, then save the document right after analyzing the layout.

If you simply want to read the document and save it, choose Read.

Once you have added all the necessary steps to your automated task and selected their properties, click Finish.

OmniPagePro

Nuance

1WaysideRoad

Burlington,MA01803

781-565-5000

Interface

DocumentManager

OfthetwoOCRprograms,OmniPageisthemoreaccessibleforsomeonewhoisblind,andthemainreasonforthatisthedocumentmanager. Asasighted-user,Igenerallyprefertoworkwithitoff,sinceIdonotwanttoallotthescreenspacetoit.ForaquickchecktoseeexactlywheretheprogramisintheOCRprocess,however,itcanbeveryuseful.

AltV+Nwilltogglethedocumentmanageronandoff.Itappearsatthebottomofthewindow,andtheareacanbemodifiedbydraggingthehorizontalbaratthetopofthepane.

Thedocumentmanageriscompletelyaccessiblewithscreenreaders(JAWS,Window-Eyes,SuperNova,etc.)andwillgiveyoualotofinformationaboutyourOmniPagedocument.Youwillneedtoconfigureit,however,orelseitwillprovidealotofconfusing,extraneousinformation.

ConfigurationforBlindUser

Thedirectionsbelowarewrittentobeusablebysomeonewhoisblindorsighted.

OmniPageincludesfourpossibleviews:

1.Thethumbnailimages:Thisviewshowsthumbnailsforeachpageandallowseasymousenavigationfromonepagetoanother.Onceinthispane,youcanalsonavigatewiththearrowkeys.

2.Thepageimage:Thisistheoriginalimageofthedocument. Boththepageimageandthumbnailimagearepartofthe“ImagePanel.”Thispanelisoptionalforablinduser.

3.Thetexteditor:ThisviewshowsthetextafterOCRhasbeenrun.

4.Thedocumentmanager:Thisviewprovidesagreatdealofinformationaboutthedocument,anditisfullyaccessibletoscreenreadersandrefreshableBrailledisplays.

AlloftheseviewscanbeturnedonandoffundertheView(Alt+V)menuonthemenubar.PleasenotethattheF6keywillmoveyoufromonepanetothenext.

Toconfigurethedocumentmanager,youneedtogoundertheViewmenutoCustomizeColumns(AltV+C).

Thedefaultatthebeginningistohaveallthecolumnson.Thisisnothelpful;infact,someofthemshowicons,whichistotallyuselessforsomeonewhoisblind.

Irecommendjustusingthefirstsixcolumnsonthelist:

Status

Characters

Suspectwords

Rejectcharacters

Words

% Accuracy

You may find that you want to add other columns later as you get used to working with the program.

Ifthescreenreaderdoesnotimmediatelystartreadingthechoicesinthechoicebox,Alt+Cshouldmovethefocustothatwindow.Thesearecheckboxes,soyoucanuseyourspacebartoselectanddeselect.

Mostly,youwillwanttodeselect.Leavethefirstfiveselectedthengothroughthelistanddeselecttherest.Thedownarrowkeywillmoveyouproperlyfromonetotheother.

Aminorpoint,butonceyouhavedeselectedtherestofthelonglist,youmightwanttouseyouruparrow(justholditdownuntilyouhear“Status”)togobacktothetopofthelist.Thecolumnsaresettoonarrowforasightedpersontoreadthecolumnlabels,andIdonotknowwhethereveryscreenreaderwillreadthemconfiguredlikethatornot.

Youcanresetthesizesasfollows:

Withthecursoronthecolumnname(forinstance,Statusisthefirstcolumnname),gototheColumnWidtheditbox(Alt+W).Typein100(theunitispixels).Alt+Cwillmoveyoubacktothecolumnlistandyoucanarrowdowntothenextcolumnnameandrepeattheprocess.

IfyouareusingaBrailledisplay,youmaywanttoplaywiththecolumnsizesabitmoretogetaprecisefit.The100pixelwidthisjustaruleofthumb.Foramoreprecisefit,trystartingwiththesesettings:

Status:67pixels

Characters:68pixels

RejectCharacters:103pixels

Words:48pixels

SuspectWords:87pixels

Onemorerefinement:Iwouldrecommendgoingtothe“RejectCharacters”choiceandmoveitup(Alt+U).Nowarrowdowntothe“Words”choiceandagainmoveitup(Alt+U).Thisorderchangeallowsyoutoheartotalcharactersfollowedbysuspectonesandthentotalwordsfollowedbysuspectwords,whichgivesyouasenseofhowwell-recognizedthepagewas.Toomanyerrors,andyoumayneedabetterscan.

MakesuretosayOKwhenallofthisisdone.OmniPagewillrememberthesettings,soyoushouldonlyneedtodoallthisonce.

IfyourfocusisnotintheDocumentManagerpane,usetheF6keytonavigatetotheproperpane.

Asthescreenreaderreadsacrosstherowinthedocumentmanager,youwillhearthepagenumberwhereyourfocusis,thenthestatus,thenthecharactersonthepage,thenumberofrejectedcharacters,thenumberofwordsonthepage,andthenumberofsuspectwords.

Youcanusethearrowkeystomovefrompagetopage.

YoumayfindthatyouneedevenfewercolumnsthanI’vesuggested,butthesefiveareanicestartingpointforyoutoexperimentwith.

Asanexampleofonewaytousethedocumentmanager,youmayhaverecognized(i.e.,runOCR)onaportionofthedocumentandthenstopped.Atalatertime,youwanttoresumeprocessingthedocument.

Tofindoutwhereyouleftoff,youwouldmovedownthelistuntilyoufindapagethathasnotbeenrecognized.Itwillsayacquired(meaningthattheimagefileispresentbutOCRhasnotyetbeenrun).

Startingatthefirstonethatsays“Acquired,”holddownyourshiftkeyanduseyourdownarrowuntilyouhaveselectedallthepagesinthedocument.

Oncethepageshavebeenselected,useControl+2toruntheOCR.

IliketoruntheOCRseparatefromtheproofreading,asIbelievethatcombiningthetwoleadstomorecrashing.

Ifthetwoarestilllinked,thatis,iftheproofreadingwindowopenswhiletheOCRisstillrunning,thenchangeyouroptions.GotoToolsandOptions:AltT+O.IntheOptionswindow,gototheProofingtab.Uncheckthesecondcheckboxonthelist,“AutomaticallyproofreadresultsafterOCR.”ThensayOK.

Toolbars

Tohelpyoubecomeaccustomedtousingtheshortcutkeys,turnontheshortcutkeysonthetoolbars:ViewToolbars.ChecktheShow ToolTips and ShowShortcutKeysin ToolTips checkboxes.

OptionsSet-up

Inthemenus,gotoToolsOptionsorusethekeyboardshortcutAlt+T+O.

OCRTab

TheOCRtaballowsyoutoselectthelanguagesthatareincludedinadocument.Ifyouareworkingonaforeignlanguagedocument,youcanchooseEnglish,aswellastheforeignlanguage.Choosingaforeignlanguageallowstheprogramtorecognizethatlettersinthatlanguagethatusediacriticalmarks(umlauts,cedillas,accents,etc.).Ifyouareprocessingamathbook,turnonGreeksothattheGreekletters(alpha,pi,theta,etc.)usedinmathematicswillberecognized.Besurethatyouonlyselectthelanguagesthatyouneed,however,astheprogramwilllookfor(andfind!)thelettersusedinthoselanguages.

Ifyouareprocessingadocumentthathasagreatdealoffinancial,legal,ormedicalterminology,besuretoturnontheprofessionaldictionaryforthatsubject.Pleasenotethatyoucanalsoloadcustomdictionariesifthereisagreatdealofunusualvocabulary.

ProcessTab

TheOCRtaballowsyoutospecifywheretoaddpagestotheOmniPagedocumentifyouhavesomehowskippedapageduringscanningorneedtoaddapagethatneededtoberescanned.MakeaTIFFfileofthepagetoaddandthensettheoptioninthiswindowpriortoinsertingit.Notethatyoucansettheoptionto“beforecurrentpage”ifyouneedtoinsertapageinthemiddleofthedocument.

Makesuretosetthe“Pagerotation”optionto“None.”Youcanmanuallyrotateapageifyouneedto,andtheprogramdoesnotdoagoodjobofanalyzingorientation.

ProofingTab

Dependingonwhetherornotyouwillrunaspellchecklaterduringyourworkflow,youmaycheckoruncheckthe“Markallnon-dictionarywords”option.IfyouwillberunningaspellcheckinMSWord,youcanleavethisuncheck.

Makesuretouncheck“AutomaticallyproofreadresultsafterOCR.”Youcanlaunchtheproofreadingwindowautomatically,andthereareothermethodsforproofreadingyourdocumentthatyoumaypreferusing.

IfyouwishtousetheIntelliTrainfeature,youwillneedtousetheproofreadingwindow.IntelliTrainremembersthecorrectionsthatyouhavemadeandlooksforpatternsthattheprogramcanapplytobetterrecognizeunusualfonts.ItmakesmostsensetouseIntelliTrainifyouareprocessingabookinsections.YoucanloadtheIntelliTraindatafromaprevioussectionbeforeprocessingthenextone,ensuringbetterrecognition.Ifyouusethisfeature,checktheboxtoprompttosave.Ifyouarenotusingthefeature,thenuncheckthatbox.

YoucanloadIntelliTraindatafromapreviousfilebyselectingthethree-dotbrowsebutton.

GeneralTab

Thereareanumberofusefulfeaturesonthistab,dependingonhowyouuseyourprogram.YoucanaddOmniPagetoyoursystemtrayforquickaccess.Youcanenableupdatedetectiontomakesurethatyoualwayshavethelatestversions.Wewillnotbeusinganyofthefeaturesinthistabforthisclass.

TextEditorTab

UnderShow,marksurethat“HeadersandFooters”isunchecked.Otherwise,OmniPagewillattempttoautomaticallydetectthesefeatures,andyoucannoteasilyaccessthem—eithertoincludethepagenumberortoreorderthezone.ItalsowillnotincludetheheaderorfooterwhenyousaveouttoWord,whichmeansthatpagenumberswillbelost.

Makesuretocheck“ReadingOrder.”Ifthisfeatureisnotchecked,youwillnotseethereaderorderofthezones(i.e.,whichzonecomesbeforewhichotherzone)orreorderthem.

Ifyourtextinthetexteditorpaneeverlooksfunny(runningoffthescreen),checktomakesurethat“WraptoRuler”isturnedon.Havingitonisthedefault,sothisshouldnotbeanissue.

ScannerTab

WedonotgenerallyrecommendusingOmniPageforscanning;however,ifyouneedtodoagreatdealofcolorscanning,youmightwishtouseif.IfyoudouseOmniPageforscanning,makesuretoinstalltheactualscanningprogramforyourscannerinOmniPage.YoucanembedthedriversthatcamewithyourscannerintheOCRprogram.DonotusethegenericdriversthatcamewithOmniPage.Evenusingthistechnique,youmayfindalossofspeedandfeaturesscanningthroughOmniPage.Unlessyouhaveacompellingreasontoscanthroughthisprogram,stickwithusingthescanningutilitythatcomeswithyourhighspeedscanner.

ProcessinganImage(TIFForPDF)File

StepOne:LoadaFile

StepTwo:RuntheOCR

BesuretoselectthepagesbeforerunningtheOCR.ClickonthefirstthumbnailanduseCTRL+Atoselectall.

StepThree:AdjustZones

Usethe"on-the-fly"tooltoredrawzones.

Toreorderzones,rightclickinTextEditorviewandchangereadingorder.Youmayneedtoungroupthezonesfirst.Rightclickandchooseungroup.

Nextrightclickandchoose,“ChangeReadingOrder.”Youwillnowseenumbersshowingtheorderofthezones.

Rightclickagainandchoose“DefineReadingOrder.”Yournumberswillnowdisappear,andyoucanclickonthezonesintheorderthatyouwishtohavethemread.

When you are satisfied with the order, right click one last time and select “Apply Order.”

StepFour:SavetheDocument

Makesuretosettheviewforyourtexteditorto“FormattedTextView”beforeyousave.ThesettingisunderViewTextEditorViewsorthebuttonsatthebottomleftofthetexteditorpane.

Whenyousavethedocument,youhaveanumberofoptions.YoucansavethetextinMSWordorothertextformats.YoucansavethedocumentasagraphicalPDForaTIFF.Youcanalsochoose“saveasmultiple”andsavetoWordandPDForWordandTIFFatthesametime.

IfyouhaveaPDFdocumentthatyouwanttoconverttoTIFF,youcanloadthefileintoOmniPageandthenjumpdirectlyto“saveas”(step3),withoutrunningtheOCR.Sinceyouareessentiallygoingfromonegraphicalformattoanother,thereisnopointtointroducingpossiblerecognitionerrorsbyrunningOCR.TheonlyreasontoruntheOCRwouldbeifyouwantedtodeleteportionsofthedocumentbeforesavingittoTIFF.

IfyouaresavingtoWordandyoudonotwantanytextboxes,youwillneedtoremovethe“retaindropcap”option.AfterselectingMicrosoftWordasthefiletypetosave,clickonthe“Options”button.Scrolldownuntilyouseethe“Retaindropcaps”checkboxanduncheckit.Notethatyoucancheckthe“Makechangespermanent”buttonifyouwanttochangethedefault.

AlsonotethatyouhaveachoiceofsavingtheentireOmniPagefiletoonedocument,savingindividualpagesasseparatedocuments,andothervariations.

OmniPageTips

Shortcut:Toseetheshortcutkeysassociatedwithmenuitems,gotoViewToolbarsandcheck"withshortcutkeys."

On-the-flyzoning:Youcannowmodifyzoneswithouthavingtoreimagetheentirepage.Forsmallchanges,clickonthe"onthefly"button.

Stopspellcheck:TotelltheprogramjusttofindOCRerrorsandnotunknownwords,gotoToolsOptionsProofinganduncheck"Marknon-dictionarywords."

Reorderingzones:Youcanreorderthezonesinthetexteditorwindow;however,itmustbesettoTruePageview(ViewTextEditorViewsTruePage).Ifthezonesarelocked,rightclickandchoosethe"ungroup"option.Thenchoose"changereadingorder"and"definereadingorder."

Verifier:Azoomwindow,calledtheverifier,isincludedinthetexteditorview.Clickonshow/hideverifieroruseF9whileinthetexteditorwindow.

Saving:Notethatyoucansaveonedocumentasindividualpages,savemultipledocumentsintoone,orsaveeachimageasonedocument.Beawareofwhichyouchoose.

Formats:ForPDF,useTruePage.ForWord,useFlowingPageorRetainFontsandParagraphs.

Batchmanager:ThebatchmanageristhenewschedulerforrunningOCR,convertingdocuments,etc.Notethatyoucanalsocancelascheduledjobthroughthebatchmanager.

Templates:Forbookswithastandardlayout,youcancreateatemplatespecificallyforthatbook.

Speech:OmniPagewillreadtexttospeechaswellasrespondingtoverbalcommandsinsomeofthewindows.

April 27, 2010

Foothill-Deanzacommunitycollegedistrict

AdvancedOCR

TableofContents

Basic Workflow

CreatingtheImageFile

AbbyyFineReader

Interface

ToolbarSet-up

OptionsSet-up

DocumentTab

1. Scan/OpenTab

2. ReadTab

Important!

3. SaveTab

ViewTab

Advanced Tab

SpellChecker Settings

ProcessinganImage(TIFForPDF)File

StepOne:OpenanImage File or a PDF File

StepTwo:AnalyzeLayout

StepThree:AdjustAreas

StepFour:ReadDocument

StepFive:CheckSpelling

StepSix:SavetheDocument

FineReaderTips

Automating Tasks

Creating an Automated Task

OmniPagePro

Interface

DocumentManager

ConfigurationforBlindUser

Toolbars

OptionsSet-up

ProcessinganImage(TIFForPDF)File

OmniPageTips