MIGRAPH OCR OPTICAL CHARACTER RECOGNITION SOFTWARE CONTENTS Welcome.................................................1 How to use this manual.............................2 What is OCR?.......................................2 The Main Steps in OCR Processing...........3 What Migraph OCR Can Do....................4 Pretrained Typefaces...............................5 System Requirements................................5 Installing Migraph OCR.............................5 Amiga Installation.........................6 Atari Installation.........................8 A Look Around the Screen................................9 Quick Start............................................11 Scanning...............................................18 Before Scanning...................................17 Document Quality..........................17 Contrast Adjustment.......................17 Alignment.................................18 Which DPI?................................19 Scanning Documents with a Hand Scanner............19 Text layout...............................20 Scanning from within Mighraph OCR.........22 Preview................................................27 Viewing an Image..................................27 - Page i - Defining Regions..................................28 Creating a Rectangular Clip Box...........28 Creating a Polyline Clip Box..............29 Defining Regions on a Full Page...........30 Altering a clip Box...............................31 Sizing the Clip Box.......................31 Combining Clip Boxes......................32 Moving a Clip Box.........................33 Deleting a Clip Box.......................33 Sorting Text Regions......................33 OCR Processing.........................................35 Automatic OCR.....................................35 Interactive Learning..............................36 Processing a File.........................36 Control Panel..........................................39 Input.............................................40 Using an Existing Image File..............40 Direct Scanning...........................40 Document..........................................41 Language..................................42 Dictionaries..............................42 Text......................................43 Interactive Learning......................43 Output............................................44 Text......................................44 Cache.............................................45 Get Image, OK and Cancel..........................46 Menu Commands..........................................47 Desk/MI_OCR.......................................47 Migraph OCR Info..........................47 Palette...................................47 File..............................................47 Load File.................................48 Save Region...............................48 - Page ii - Default Settings..........................48 Load......................................48 Save......................................48 Save As...................................49 Current...................................49 Quit......................................49 Scan..............................................50 Start Scan................................50 Settings..................................50 Clip..............................................51 Undo......................................51 Delete....................................51 Select All................................51 Sort Order................................51 Erase Graphic.............................52 # Text Regions............................52 Process...........................................52 Control Panel.............................52 OCR.......................................52 Rotate....................................53 Invert....................................53 Flip......................................53 View..............................................53 Zoom Full.................................53 100%......................................54 200%......................................54 Show Order................................54 Locator...........................................54 Index..................................................55 - Page iii - WELCOME Migraph, Inc. is proud to present you with Migraph OCR (Optical Character Recognition), a powerful, intelligent, full featured text reading application. Migraph OCR offers one of the fastest and least expensive means of converting printed text into an Ascii text file which can be loaded into word processing and desktop publishing programs. While it is convenient to scan directly from within Migraph OCR using one of the compatible hand scanners, you do not have to own a scanner to use Migraph OCR, you need only have access to one. Since Migraph OCR will load scanned TIFF, IMG (Atari), and IFF (Amiga) files, you may use any scanner that saves these file formats. You may even use a scanner on another computer (a PC for instance), as long as it saves one of the compatible monochrome formats. You will find Migraph OCR easy to use. Optical Character Recognition is one of those applications where all the hard work is done behind the scenes. User interaction is normally minimal, depending on the quality of the scanned text being processed. Like other Migraph programs, Migraph OCR is user friendly, employing icons, dropdown menus, and dialog boxes. Thank you for purchasing Migraph OCR. We know you'll find it to be one of your most useful applications. Before installing this program, please take the time to fill out the registration card and mail it in. Being a registered owner entitles you to receive free technical support and the Migraph newsletter "Headlines", which provides news on updates, upgrades, new products, etc. - Page 1 - HOW TO USE THIS MANUAL This manual is for use on both Amiga and Atari systems. Except where noted, all information applies to both systems. The manual is divided into several chapters. After reading this chapter and installing Migraph OCR, got to the Quick Start chapter which provides a good introduction to the OCR process and lets you see how it works. The remaining chapters have detailed information on scanning, preview, and OCR processing. Hints are found throughout the manual in the appropriate sections. Note: If you own one of the compatible hand scanners, we recommend that you read the section on direct scanning before starting. WHAT IS OCR? OCR stands for Optical Character Recognition, a process whereby printed text is electronically converted into computer readable files i.e. an Ascii file. Rather than typing text into the computer by hand, an optical scanner is used to scan the text. The scanner's job is to convert the image into a map of black and white dots (pixels) that represent the original image. When scanned text appears on your monitor screen, an "A" looks like an "A", but to the computer it is just a series of black pixels grouped together. OCR software is the vital link that identifies the group of pixels as a particular character. In a simplistic view, the program basically says, this is an "a", this is a "b" and so on. Previously, OCR programs were raster based. This meant that the program relied upon a database of stored pixel images to identify the scanned text. Unfortunately, if the scanned character did not exactly match the size and shape of a character in the database, it was not recognized. The next step in OCR was a trainable OCR. With a trainable program, when an unrecognized character appeared, you could correctly identify the character for the program and thereby "train" it to recognize the character - Page 2 - in the future. Depending on the quality of the OCR program though, you might have to train a program on a character 5, 10, 15 times or more before it would recognize the character automatically. Then came Omnifont technology. Omnifont technology uses a mathematical definition to identify the character instead of a raster image. Not only is it quicker and with a higher accuracy rate, but the program itself can be considerably smaller in size. Using mathematical definitions also means that the program has a greater chance of recognizing similar typefaces, even though it may not have been trained on them. THE MAIN STEPS IN OCR PROCESSING There are four main steps in OCR processing: 1. Line segmentation 2. Word and character segmentation 3. Character recognition. 4. Text output Line Segmentation consists of slicing a page into its different lines. Line skew and interspacing are also analyzed in this step. The most difficult documents are those with a small interline spacing and a significant skew. Word and Character Segmentation isolates one word from another and then seperates each letter from another. This step is easier when the text is monspaced, meaning all the letters are the same width. With proportional text where each letter's width is based on its shape, the process takes a little longer. The most difficult case is where characters are of a proportional font and are randomly touching. Character Recognition involves extracting the characteristics of each isolated shape and giving it a name. Text Output is the conversion of the recognized text into a file which can be imported into your word processor or desktop publishing application. - Page 3 - WHAT MIGRAPH OCR CAN DO Migraph OCR is pretrained on twenty typefaces and can be trained to recognize many more. It recognizes both American and European character sets as well as proprtional and monspaced fonts, typeset, typewritten, laser printed, and NLQ dot matrix quality text. Migraph OCR can read type-written and non-kerned proportional italics fonts. Other italics may be recognized with a little slower validation process and may require more user interaction. The character size range is 10-18 points which covers most of the font sizes used for body text. However smaller sizes can be recognized when scanned at 400 dpi. Characters of mixed sizes don't pose a problem as long as there is no possibility of confusing a capital of one size with the lower case of another size. If there is the possibility of confusion, the Migraph OCR decision regarding uppercase/lowercase may be incorrect. Using Migraph OCR for titles is not recommended since they are usually in large point sizes and may be easily added in a word processor or desktop publishing program (DTP). In addition to having Omnifont technology, Migraph OCR has intelligence. Lexicons for four languages are provided: English, French, German, and Dutch. Lexicons are linguistic databases which provide additional information, like syntax, to aid in character recognition. By using the information in the lexicons, the program is actually able to make decisions regarding characters, rather than just presenting them to you for identification. While Migraph OCR is a powerful program, it cannot do everything. It will not read parts of text in dense or kerned italics typefaces and it may have trouble reading scans that are too light or too dark, or that are skewed. - Page 4 - PRETRAINED TYPEFACES Migraph OCR has been pretrained on these typefaces: Artisian, Bookman, Brougham, Caroll Pica, Courier, Courier Italic, Delegate, Elite Modern, Helvetica, Herald Elite, Letter Gothic, Lori, Lubalin, OCRB, Pica, Prestige, Elite, Prestige Italic, Prestige Pica, Times, Titan, and Title. SYSTEM REQUIRMENTS Migraph OCR requires at least 2MB RAM and a hard disk. As with any program that scans graphics, the more memory you have, the better. A full page scan at 300 dpi normally takes up about 1MB. Migraph OCR uses hard drive caching, which means that when you run out of RAM, the remaining file information will be sent to your hard disk. While this enables you to work with larger files, it is slower than using RAM memory. - Page 5 - INSTALLING MIGRAPH OCR Note: Before installing Migraph OCR, insure that you have at least 2MB available on one of your hard drive partitions. Amiga Installation To install Migraph OCR on an Amiga system, follow these directions: [1] Boot-up your computer, if it is not already on. [2] Insert the Migraph OCR Master Disk into your floppy drive. [3] Double-click on the Migraph_OCR floppy disk icon. [4] Double-click on the Install_OCR icon. Follow the prompts that appear on your screen. When done, Migraph OCR will be installed on your hard drive and ready to run. To run Migraph OCR, double-click on the Migraph_OCR icon within the Migraph_OCR drawer. Running Migraph OCR from the CLI To start Migraph OCR from the CLI, type the following line at the CLI prompt: Run Migraph_OCR Adjusting the Screen Palette Migraph OCR for the Amiga, has a palette tool which allows you to modify - Page 6 - its screen colors by altering their RGB values. This is useful for reducing screen flicker when you want to run Interlace mode. When running Migraph OCR on a custom screen, the palette alters only Migraph OCR's colors. However, when you run Migraph OCR as a Workbench application, the palette affects the colors of the whole Workbench. This is important to remember, especially should you decide to save the colors as the default. To use the palette tool, follow these directions: [1] Select Palette from the MI_OCR dropdown menu. A dialog box will appear. If the value for planes is set to two, four colors appear; if the planes value is set at one, two colors will appear. The palette will display whatever colors you have saved as the default. However, the very first time you use the palette, it will display the default colors for the workbench: black, white, blue, and grey (2.x) or blue, black, orange, and white (1.3). [2] Select the color whose RGB values you wish to alter by clicking on it. To alter the RGB values move the slider for each value. As you move the slider, the screen and window colors will change. To reduce screen flicker in Interlace mode, choose colors with little contrast between them. For instance, instead of black and white, you might choose grey and black, or red and black. It will be immediately apparent when altering the colors, whether the flicker is getting better or worse. You may need to change the RGB value of each plane color to get maximum results. If, after changing the RGB values, you decide you would prefer the original place values, click on DEFAULT. The palette will then display the original workbench colors. To reset the values back to the previously selected palette, click on Reset. [3] After the values have been changed, you may click OK to save - Page 7 - the values for the current sesion, or click Cancel to quit the palette and reset the values as they originally were before you entered the palette dialog box. [4] If you have selected a new set of values that you would like to keep as the default from the Touch-Up custom screen, select Defaults from the File menu. The Defaults dialog box will appear. Click on Save. Thereafter, these colors will be used every time you run Touch-Up. - Page 8 - A LOOK AROUND THE SCREEN Before using the tutorial in the next chapter, you may want to familiarize yourself with the elements of Migrap OCR's screen. This section provides a brief description of each of the elements found on the screen. Boot-up your computer (if you have not already done so), then start Migraph OCR by double-clicking on the Migraph OCR program icon. You will initially see a dialog box with the copyright information, and then the program screen will appear which looks like this: - Page 9 - Close Box - Use the close box to quit Migraph OCR. Menus - These menus offer access to all the commands you need to run Migraph OCR. Slider Bars - The slider bars located on the right and bottom of the screen allow you to access all parts of an image that has been scanned or imported into the program. Top Icon Pad - This pad contains five icons which are used in the preview mode before OCR processing begins. They allow you to designate text areas using boxes or polygons which can then be edited. You may also select a graphic area you wish to save. The selected icon is inverted. Bottom Icon Pad - This icon pad contains the tools and commands you'll use to import or scan images and perform OCR on them. The center icon flips between a scanner and disk depending on whether you have opted to load a file or scan directly. Zoom Levels - Migraph OCR provides three zoom levels at which to view an image, while the black box represents the total area of the scanned or imported image, while the black box represents the portion of the image displayed on the screen. To use the Locator move the mouse over the black box, and drag the box to a new location, then release the mouse button. A new part of the image will then be displayed. - Page 10 - QUICK START The easiest way to become familiar with a program is to begin using it. This chapter leads you through a step by step tutorial using a sample scanned file which can be found on your Master Disk. Even if you own a hand scanner that allows you to scan directly, we recommend that you follow this tutorial to help you understand how the OCR process works. Note: The directions in this chapter assume that you have already installed Migraph OCR on your hard drive. [1] Double-click on the Migraph OCR program icon. The program will run and the following screen will appear. - Page 11 - [2] Click on the Control Panel icon. The following Dialog box will appear: The Control Panel is the dialog box that you will use to define all the parameters for the program. You can see that it lets you specify input, output, document characteristics, dictionaries, graphic formats, and cache path settings. [3] Under Input, click on the File button (if it is not already selected). [4] Under Document, click on New under the Dictionary button. This specifies that you will be creating a new font directory. [5] Click on the Dictionary button. The File Selector will appear. Enter the path and name for the dictionary, for instance TEMP.ALD. - Page 12 - This dictionary will hold any new characters that you train Migraph OCR to recognize during this tutorial. When you process a file, the program will use any dictionary you specify, in this case TEMP.ALD, in addition to the built-in dictionary. [6] Under Output, click on Text Out. The File Selector will appear. Enter the path and file name you wish to save the sample under, for example TEST.TXT. If you do not specify a name, the Ascii file will be automatically named MIOCR.TXT. [7] Click on the Get Image button. The file slecetor will appear. Select the file DEMO.IFF and click OK. The file will load ad be displayed on your screen. [8] Click on the Zoom Full (F) icon. [9] Click on the Text Icon. This designates the area you will be defining as a text region. [10] Click on the Clip Box icon. Move the cursor to the upper left hand corner of the top paragraph and click the left mouse button once. Move the mouse until the clip box is the desired size and click the left mouse button again. This creates a clip box around the text. - Page 13 - [11] Click on the Point Mode icon. Point mode allows you to select, size and move clip boxes. To select a clip box, click inside it. To select more than one region, click inside any additional clip boxes you wish to select. To deselect any clip box, click in it again. To size a clip box, first select it, then move the mouse over one of the polymarkers, depress the left mouse button and move the mouse (also known as 'dragging' the mouse). When the clip box is the correct size, release the mouse button. To move a clip box, first select it, then place the mouse within the box, and drag the mouse. When the box is in the desired location, release the mouse button. [12] Click on the OCR icon. Several status bars will be displayed on the screen as Migraph OCR processes the image. After a short while, a dialog box will appear that looks like this: - Page 14 - This is the Interactive Learning dialog box which is the heart of Migraph OCR. This is where you will make the decision to accept or reject characters, and where you will train the program on new characters. When it processs text, Migraph OCR goes through several decision phases where it compares the characters in the file to the trained characters in the built-in dictionary and the user dictionary (if any) and tries to identify every character. When it has recognized every character on its own that it can, it will then present the unrecognized characters to you for identification. BOX #1 displays the text that has been processed (i.e. recognized). When Migraph OCR comes across a character it cannot recognize, the word that contains it will be inverted. The program will now stop and wait for action from you. BOX #2 displays the word which contains the unrecognized character. The unrecognized character will be in bold on monochrome systems, or the opposite color on color systems. BOX #3 displays the program's best guess at the correct identity of the character. If a single letter guess is incorrect, you can simply type the correct character. If theer should actually be several characters, use the backspace key to delete the current character and the necessary amount of leading spaces and then enter the correct character(s). After a correct guess or a correction, you will use one of the buttons below to specify an action regarding the character. Each of the buttons offers you a different choice as to how to handle the character in question. Here is a brief explanation of the choices. TRAIN: Use thsi option to train the program on new characters. The program will form a mathematical definition for the character and store it in the specified user dictionary. DO NOT use this option for defaced or ill formed characters that you would not want in the dictionary. Note: When the box next to the Accept ot Train button is selected (black), it is the default button. As the default button, you may use the Return key - Page 15 - instead if using the mouse. Click on your choice to select it. ACCEPT: Use Accept when you wish to accept/correct the character in box #3 and have that character placed in the file, but NOT in the dictionary. The program will not be trained on the accepted character. CONFIRM: In some fonts, the letter i and the number 1 are identical, as are the letters o and the number 0. When this is the case, you can use the confirm key to correctly specify which is which. Thereafter, the program will not present these characters to you for identification. UNDO: Undo allows you to go back and change choices you made regarding the last fifteen (15) unrecognized characters. DELETE: Use Delete to delete unwanted characters, for example dirt specks that were picked up by the scanner. AUTO: This option processes the file or remainder of the file without interactive learning. The program replaces unrecognized characters in the text file with its best guess. If it cannot make a guess (such as when characters are joined) it will replace the unrecognized character(s) with the @ sign. STOP: Use the Stop button when you wish to stop processing before the file is done. The program will save whatever text has been processed and return you to the previous screen. [13] Process the file trying different choices. After the image is processed, the text will be saved to the file you specified earlier, and you will be returned to the main screen. [14] To view/edit the text file, ext Migraph OCR by selecting Quit from the File menu; then run your word processor and load the new text file. - Page 16 - SCANNING Note: This chapter contains information on scanning that is useful to all scanner owners. Regardless of which type of scanner you own, we recommend that you read the general sections at the beggining of this chapter before scanning text for processing. BEFORE SCANNING This section contains guidelines for scanning documents. Document Quality One of the first things to look at when considering a document for OCR processing is the quality of the printed text in the document. Is it sharp, clear, dark print, easily readable? Or is it poor quality, blurry letters, hard to read? The quality of the original affects how accurate Migraph OCR will be when processing the scan of the document. Always try to have the best quality original possible. Hint: If the original is particularly light, make a photocopy and scan the photocopy. Contrast Adjustment Scanner contrast is one of the most important factors which can affect recognition accuracy. Scans that are either too light or too dark may cause - Page 17 - the OCR to recognize fewer characters, which in turn means more interaction by you to determine character identity. Scans that are too light may cause parts of characters to drop out. For instance an "o" may look like (). Scans that are too dark cause the letters to thicken, and the openings in characters like "a" and "e" tend to close up. The space between characters also decreases making it more difficult for the program to discern individual characters. Using the contrast control on your scanner allows you to increase the quality of poor originals, and to get the best quality from good originals. Since the quality of documents does vary, we recommend that your experiment with the contrast level before you process the scan. If the contrast is too light or dark, rescan the image before processing. Alignment After contrast, the next most important item is alignment of the document to the scanner. If the scanned image is skewed, it will affect the accuracy of the OCR because characters from the first line may cross over and enter the area of the second line, etc. This interfers with the OCR's processing step of line segmentation. - Page 18 - Should your scan be skewed, we recommend that you rescan the image. If you are using a hand scanner, try using a book as a straight edge, or a scanning device, like the Migraph Scanning Tray, to get a straighter scan. Which DPI? When you wish to scan text between 10-18 points, we recommend that you scan at 300 dots per inch. For text less than 10 point we recommend you scan at 400 DPI, as this will provide more dots and a better image. For text over 18pt, use 200 DPI. SCANNING DOCUMENTS WITH A HAND SCANNER You may scan text and graphics directly from within Migraph OCR using the Migraph, AlphaData, and Golden Image hand scanners. Direct scanning is preferable because it saves time since you will not have to spend time saving and loading files that will only be deleted later. If you own one of the compatible hand scanners, we recommend that you always scan directly from within Migraph OCR. If you must scan outside of Migraph OCR with a hand scanner, scan the page in several passes, saving each pass out as a seperate file. While it is helpful to use a commercial scanning tray to get a straight scan, it is not - Page 19 - necessary to merge the passes into one file. Merging the files will not necessarily speed the processing and will take up more of your time and disk space. Note: Through the remainder of the chapter, it will be assumed that you are scanning from within Migraph OCR. If you are not, the only additional steps would be to save the scan as an IFF, IMG, or TIFF file and to load it into Migraph OCR for processing. Otherwise the instructions are identical. Text Layout How text is placed on a page and how much of it you want to scan will determine whether you scan horizontally or vertically, and how many passes you use to scan a page. If you only need to scan one column of text or a few paragraphs, you can simply scan the text and process it. However, to scan a full page of text, you must do it in several passes. The basic process of scanning a full page of text with a hand scanner, after setting all the parameters in the Control Panel, is as follows: [1] Scan the first column or portion. [2] Define the text region you wish to process with a clip box, excluding any text you do not want (such as any overlap from the second column). [3] Process the text. It will be saved to the new file you specified. [4] Re-enter the Control Panel, select Append under Output, and click OK. Thereafter, any additional text you process will be appended to that file. [5] Repeat steps 1-3 for any additional text columns/pages until the document is completely processed. - Page 20 - Hint: When you know you will be appending text, just select Append to start with when you name a file. When Migraph OCR goes to save the text to the file and does not find it, it will create the file automatically. The following illustration shows scanning the image, the image itself, and the defined text region from the scan that will be processed. Migraph OCR appends text in the order you process it. For example, on a page with three columns, you would always scan and process the first column first, the second column second, and the third column last. If you do not go in this order, the final text will not be correct. Hint: If you are scanning outside of Migraph OCR, do all the scans first, then load/process the files in order. If you have scanned the image horizontally, load the image, then select Rotate Left or Right under the Process menu to rotate the file to the correct orientation. The following illustration shows the two basic page layouts. The recommended scanning area is shows with dashed lines and the scanner shows the scanning orientation (horizontal or vertical). The numbers represent the order in which the columns should be scanned/processed. - Page 21 - Hint: When scanning horizontally, use a white sheet of paper to cover the area you do not want to scan so that you have a clean edge. You could also use this tip when scanning vertically to avoid any overlap from another column. Scanning from within Migraph OCR These instructions assume that you are in Migraph OCR, have your hand scanner properly installed, and know how to scan an image. [1] Prepare the document you wish to scan by making sure it is straight. Use some type of straight edge or guide for the scanner head to insure a straight scan. [2] Select the line art setting on the scanner head. [3] Adjust the contrast wheel if necessary. [4] Click on the Control Panel Icon. - Page 22 - The control Panel will appear. It look like this: [5] Under input, click on Scanner (if not already selected). [6] Click on the Settings button. The Scan Area dialog box will appear. - Page 23 - [7] Select the DPI you will be scanning at, by clicking on it. At the same time, make sure the DPI setting on your scanner head is set at the same value. (For this example, use the default of 300 DPI.) [8] Set the scan length by either moving the sldier bar, or by clicking on the arrows or the grey slider box. The value is displayed to the left of the slider area and will change as you move the slider. (For this example, set a length of 4 inches.) [9] Vertical scanning is the default and should already be selected. If it is not, select it by clicking on the vertical scanner icon; then click OK to exit the dialog box. [10] Under Output, click on the New button. This will create a new file that the Ascii text will be saved to. Click on the Text Out button. The File Selector will appear, select a path and name for the file and click OK. Not: If you know you will be processing several scans for one text file, select Append rather than New. This eliminates having to reenter the Control Panel. When Migraph OCR does not find the file you have named after processing the image, it will automatically create the file. [11] Click on the Get Image button. An alert will appear that states the scan will destroy the current image. Click OK to begin scanning. The scanning light will come on. Scan the image as you normally would and release the start button when done scanning. If you decide not to scan or if you release the button before scanning the full length, the scanning light will automatically turn off after 10 seconds. The scanned image will then appear on your screen. Note: Scan slowly, making sure the speed indicator light does not blink. Blinking indicates a possible loss of data that could hinder the OCR process. - Page 24 - [12] Use the locator to view other parts of the image. Move the mouse to the black box in the Locator, then drag the black box to a new area and release the mouse button. Look to see that the characters have good contrast, being neither too light nor too dark. If the contrast is not correct, rescan the image. Check the image to make sure it is not skewed. If it is, rescan it. [14] Click on the Zoom Full icon Once an image has been scanned to your satisfaction, you may select the text regions you wish to process or a graphic region you wish to save. If the graphic is a halftone, you may want to rescan the graphic portion of the image using one of the dither patterns on the scanning head. Note: Should you forget to access the Scan Area dialog box through the Input button, it will automatically appear when you select the Get Image button the first time you try to scan in a session. Once you have used this dialog box, you do not need to enter it again unless you wish to change the length or direction of the scan. - Page 25 - NOTES: - Page 26 - PREVIEW Preview is the time between scanning/loading a file and OCR processing. During this stage you may view the image and select the text regions you wish to process and any graphic regions you wish to save. A clip box is used to define these regions. While you may have multiple text regions defined on a page, you may only define one graphic region at a time. Note: If you have a scan which would normally require only a single clip box around the entire page, you do not have to define text regions. For example a typewritten letter which is in paragraphs, rather than columns. When you begin the OCR process, you will be asked if you want to process the entire page. Remember, if there is any text which was clipped by the edges of the scan, you should first define the text using a clip box. VIEWING AN IMAGE Once an image has been scanned or loaded, you may wish to view it at different zoom levels to check for skew, contrast, etc. The image will automatically be displayed at Zoom Normal. To view other parts of the image, you may use the Locator or the slider bars to the right and bottom of the image. To see a more enlarged view then Zoom 1, select one of the other zoom levels. To view the whole image, click on the Zoom Full button. Zoom ull is handy because it displays the whole scan/page making it easier to define regions. - Page 27 - DEFINING REGIONS Text regions may be defined using either rectangular or polyline clip boxes. Polyline clip boxes are very useful for irregular areas, such as when text flows around a graphic. Once graphic region may be defined/saved at a time using a rectangular clip box. Creating a Rectangular Clip Box To create a rectangle clip box, follow these directions: [1] Click on the Text icon if you will be defining a text region. OR Click on the Graphic icon if you will be defining a graphic region. [2] Click on the Clip Box icon [3] Move the mouse to the upper left corner of the region you want to define and click the left mouse button once. Move the mouse until the clip box is the desired size and click the left button again. - Page 28 - Creating a Polyline Clip Box A polyline box is used when you need to define irregular shaped text areas. Often a graphic may be located within a tet column or have text wrapped partially around it. In these cases a rectangular clip box would either not capture all of the text or include part of the graphic with the text, neither of which is satisfactory. Since the polyline clip boxes with Migraph OCR must have all line segments at 90 degree angles, it's important to think about the shape you are going to use beforehand. To create a polyline clip box, follow these instructions: [1] Select the text icon. [2] Click on the Polyline icon. [3] Move the mouse to the area you wish to define. It is usually a good idea to start at the top left of the region. [4] Click the left mouse button to start the segment. As you move the mouse a line will be drawn to the left/right of the starting point depending on where you move the mouse. The first seg- - Page 29 - ment is always horizontal. When the line is the correct size and in the correct location, click the left mouse button again. After specifying the end of a segment, the next segment will always turn at a 90 degree angle. [5] Continue step 4 until the polyline is one segment short of being complete. To close the polyline double-click the left mouse button and the last line segment will appear automatically. Note: After the first line segment is drawn, you may move the mouse in any direction in order to extend the line or to turn 90 degrees. You may not go back in the same direction you just came from. Note: To abort a polyline clip box while creating it, hit the escape key. Defining Regions on a Full Page You may define multiple text regions within Migraph OCR. The number of regions you have currently defined is displayed in the Clip menu. When a full page image is loaded, you will be able to define the text regions for the whole page before processing it. Because Migraph OCR processes text regions in the order they are defined/numbered, remember to select them in the order you wish to have them appear in the final text file or to - Page 30 - resort them to prior processing. Otherwise, the text file will not be in the same order as the original document and you will need to edit the text file. ALTERING A CLIP BOX Once a clip box has been created you may resize, move, or delete it. To alter a clip box, follow these directions: [1] Click on the Point Mode icon. [2] Select a clip box by clicking on it. To deselect it, click on it again. A selected clip box will display polymarkers (small boxes) around its border. Note: A clip box must be selected before you can size or move it. Sizing the Clip Box To size the clip box, move the mouse over a polymarker, and drag the mouse. Whne the clip box is the correct size, release the left mouse button. The clip box will grow or shrink depending on the direction you moved the mouse. - Page 31 - When sizing a text clip box, if it should overlap another defined text region, the two clip boxes will combine and form one region. Depending on the page layout, this may not be desirable. An undesirable instance is when two full length vertical text columns are defined and they join to become one. Because Migraph OCR first segments the text by lines, it will read the top line of the first column and then the top line of the second column, and so on. This is how the text will be displayed in the processing dialog box and the final text file. Editing would then take place in a word processor. Combining Clip Boxes Rectangular clip boxes can be used to define irregular regions. You may use several overlapping rectangular boxes to quickly define a complex region. In fact, it is often quicker and easier than using a polyline clip box. To define an irregular area using rectangular clip boxes, follow these directions: [1] Select the clip box tool and define the first region. [2] Still using the clip box tool, define the remaining regions, making sure to overlap them where needed. The clip boxes will join automatically. - Page 32 - Moving a Clip Box To move a clipbox, place the mouse inside a selected clip box, then drag the mouse. A hand appears in the box while you are moving it. Move the box to the new location and release the left mouse button. Deleting a Clip Box To delete a clip box, first select the clip box, then either choose Delete from the Clip menu or use the Delete key. After deleting a clip box, use the Show Oredr command to make sure the text regions are in the correct order for OCR processing. If they are not, use the Sort Order command to resort them. Undo The Undo command located in the Clip menu will undo the previous clip box command. Undo can also be used to undo an undo (for those times when you reversed a command and then decided you didn't really want it reversed). SORTING TEXT REGIONS Migraph OCR processes multiple text regions in the order they were defined. Because of this, there will be times when text regions need to be sorted, otherwise, the text in the Ascii file will appear in a different order than the original document. For example, Sort Order would be used when a text region has been deleted *interferring with the order), or when the regions were defined out of order. The Sort Order command, found under the Clip menu, allows you to define (or redefine) the order of the text regions for OCR processing. The Sort Order command is generally used for full page images where more than one or two text regions can be defined. Hand scanner owners should never really have need of this feature. - Page 33 - Note: If you have defined the text regions in the correct order, there is no need to use the Sort Order command. To sort defined text regions, follow these drections: [1] Select Sort Order from the Clip menu. The view will change to Zoom full, and the existsing order will be removed. All text boxes will be displayed and any graphic region hidden. The Show Order command is automatically activated so the order is displayed on the screen. [2] Select text regions one at a time, starting with the region you wish to have processed first. From the second region n, the order will be displayed. A line will be drawn from region to region in the order in which they were selected. Each region is assigned a processing number which appears above the upper left hand corner of the region. [3] When you have selected all but the last text region, the new order will be completed and you wil be out of the Sort Order command. Should you decide to change the order, simply select the Sort Order command and start over. To remove the display after sorting, deselect the Show Order command under the View menu. Note: To abort the Sort Order command, press the Escape key. If you wish to abort, you must hit escape before selecting the next to last region. Note: Text layouts with only one column of text do not need to be sorted, whether scanned by a flatbed or hand held scanner. - Page 34 - OCR PROCESSING After loading an image file into Migraph OCR (whether by scanning or loading an existsing file), and selecting the different text regions, you are now ready to run OCR on the image. This chapter explains how the OCR process in this program works, and how to use the Interactive Learning mode. AUTOMATIC OCR Two types of OCR processing are available in Migraph OCR: Automatic and Interactive Learning. Automatic OCR processes the file without any interaction by you. It is quicker than Interactive Learning because it does not require your interaction on questionable characters, but no learning or correct to doubtful characters takes place. When in Automatic mode, the program replaces unrecognized characters with its best guess. If it has no guess, for example when two characters are attached, it replaces the unrecognized character(s) with an @ symbol. Any corrections to the file would then take place in your word processor or desktop publishing program. To use Automatic mode, enter the Control Panel and turn Interactive Learning "off" by deselecting it. Note: Automatic mode would only be used when the current user dictionary has been trained on the same font styles as the current image. - Page 35 - INTERACTIVE LEARNING Interactive Learning is most often the better choice for OCR processing. While it is the slower of the two methods, it provides you with the opportunity to identify unrecognized characters. To use Interactive Learning, enter the Control Panel and click on the Interactive Learning box. Note that it is the default, so if you haven't changed the setting to Automatic, you need not change anything. Processing a File When OCR is performed on an image, whether Automatic or Interactive, the program goes through multiple decision making phases behind the scenes before presenting its solutions to you. It will run through the file several times trying to identify the characters from those in its main dictionary as well as any dictionary you have specified. It will also use the lexicons to help idetify unrecognized characters and to formulate the best possible guess as to the identity of each character. Once this phase is over, the Interactive Learning dialog box will appear and present the unrecognized characters to you. The Interactive Learning dialog box looks like this: - Page 36 - BOX #1 displays the text that has been processed (i.e. recognized). When Migraph OCR comes across a character it cannot recognize, the word that contains the character will be inverted. The program will now stop and wait for action from you. BOX #2 displays the word which contains the unrecognized character. The unrecognized character will be in bold on monochrome systems, or the opposite color on color systems. BOX #3 displays the program's best guess at the correct identity of the character. If a single letter guess is incorrect, you can simply type the correct character. If there should actually be several characters, use the backspace key to delete the current character and the necessary amount of leading spaces and then enter the correct character(s). After a correct guess or a correction, you will use one of the buttons below to specify an action regarding the character. Each of the buttons offers you a different choice as to how to handle the character in question. Here is an explaination of the choices. TRAIN: Use this option to train the program on new characters. The program will form a mathematical definition for the character and store it in the specified user defined dictionary. DO NOT use this option for defaced or ill formed characters that you would not want in the dictionary. Note: When the box next to the Accept (or Train) button is selected (black), it is the default button. As the default button, you may use the Return key instead of clicking on the button. To change the default, click on the other button. ACCEPT: Use Accept when you wish to accept/correct the character in box #3 and have that character placed in the file, but NOT in the dictionary. The program will not be trained on the accepted character. CONFIRM: In some fonts, the letter i and the number 1 are identical, as are the letter o and the number 0. When this is the case, you can use the confirm key to correctly specify which is which. Thereafter, the program - Page 37 - will not present these characters to you for identification. UNDO: Undo allows you to go back and change choices you made regarding the last fifteen (15) unrecognized characters. DELETE: Use Delete to delete unwanted characters, for example dirt specks that were picked up by the scanner. AUTO: This option processes the file or remainder of the file without interactive learning. The program replaces unrecognized characters in the text file with its best guess. If it cannot make a guess (such as when characters are joined) it will replace the unrecognized character(s) with the @ sign. STOP: Use the Stop button when you wish to stop processing before the file is done. The program will save whatever text has been processed and return you to the previous screen. To view/edit the final text file after processing, load the text file into a text editor, word processor, or DTP program. Hint: Using a text editor or word processor which operates as a desk accessory will save time when you wish to view the file. Note: Additional information regarding training can be found in the chapter on the Control Panel under Dictionary. - Page 38 - CONTROL PANEL The Control Panel is where you will select all the settings that govern image input and output, file types, document specifications, and more. This chapter describes the features of the Control panel in depth and how to use them. The Control Panel is brought up by either clicking on the tool icon or by selecting Control Panel from the process menu. The Control Panel looks like this: The Control Panel is divided into four sections: Input, Document, Output, and Cache. - Page 39 - INPUT The Input section specifies whether the image (input) will be directly from a scanner or from a previously scanned image file. Using an Existing Image File When you wish to load an existing image file, select File by clicking on it. The File Mask button allows you to select the file type you wish to load, and displays only that file format in the File Selector. The currently selected file type is displayed to the right of the button. To change the file type, click on the File Mask button. When the pop-up menu appears, make your choice by clicking on it. Note: You may load either TIFF or IMG (Atari) or IFF (Amiga) files from the File menu, regardless of the File Mask set in the control Panel. Direct Scanning If you have one of the campatible hand scanners (and enough memory) we strongly recommend that you scan from within Migraph OCR. To do this, select Scanner. Before the first time you scan in a session, the program expects you to set certain scanner parameters. To do this, click on the Settings button. The Scan Area dialog box will appear. - Page 40 - Within this box you will set the dots per inch (dpi), the scan length, the unit of measurement, and whether you wish to scan vertically or horizontally. DPI - To select a dpi, click on your choice, making sure it corresponds to the dpi set on the scanner head. The default is 300 dpi. SCAN LENGTH - The scan length value is displayed in the upper left corner of the length section. This value is based upon the amount of memory avaialble at the time Migraph OCR starts up. This value can only be changed by using the slider bar. You cannot edit the value itself. To setthe scan length, you may move either the white slider bar, click on the arrows, or click in the grey area. If you click in the grey area, one click moves the slider by .1 cm, or .1 inch, depending on which unit of measurement you have selected. UNIT OF MEASUREMENT - To select a unit of measurement, click on your choice, either inches or centimeters. The value of the scan length will correspond to the selected unit of measurement. SCANNER ORIENTATION - Migraph OCR allows you to scan both vertically or horizontally. The default is portrait (vertical). To scan horizontally, click on the landscape scanner icon. When you scan an image horizontally, Migraph OCR automatically rotates the scan so that it appears vertically on your screen. After making your choices, click OK to return to the Control Panel. Note: When scanning horizontally, scan from left to right. DOCUMENT The Document section provides Migraph OCR with specific information regarding the document you wish to process. You will specify the language, the user dictionary you wish to use, text size, text pitch and whether you wish to use Interactive Learning or not. - Page 41 - Language Migraph OCR includes lexicons for English, French, German, and Dutch languages, with English being the default. The current language is displayed to the right of the Language button. To select a different language, click on the Language button. When the pop-up menu appears, make your choice by clicking on it. Dictionaries Migraph OCR allows you to create user defined dictionaries in addition to the built-in library. In this case, dictionary does not refer to spelling, but instead to a collection of stored character definitions. Any characters you wish to train the program on will be stored in the user defined dictionaries. Before processing a document, you may choose to create a new user dictionary, append an existing user dictionary, or use only the built-in dictionary. User dictionaries can be quite handy, especially when you have a multipage document that may have special characters like math symbols or Greek characters. Differenrt dictionaries can be created for specific types of documents. NEW - Select New when you wish to start a new user dictionary. Click on the Dictionary button. The File Selector will appear. Enter the name for the new dictionary and click OK. Dictionaries have the file extension .ALD. Shoudl you forget to add the file extension, Migraph OCR will automatically add it for you. APPEND - Select Append when you wish to use the existing characters in a user dictionary as well as add new caharcters. READ - Select Read when you wish to have the program use the existing characters in a user dictionary, but not learn any new ones. When adding new characters to a user dictionary, make sure they are char- - Page 42 - acters you will want to keep. We recommend that you do not train the program to recognize defaced or damaged characters, as they will be used infrequently and take up space in the dictionary file. The exception to this is when you have a large document or several documents that are of poor quality and will not scan well. In this case, you may want to have a special dictionary that contains defaced or irregular characters that will be used when processing these documents. WE STRONGLY RECOMMEND THAT YOU CREATE DIFFERENT USER DEFINED DICTIONARIES FOR DIFFERENT STYLES OF DOCUMENTS. For instance, you might create one for typewritten documents, one for laser printed documents, one for typeset documents, etc. We have found that the larger the dictionary is, the slower the OCR processing will be. Another reason for having separate dictionaries is that during the interactive learning process, sometims the least used characters in a dictionary will be eliminated in favor of more frequently used characters. Text Document text is classified as either small (under 10 point) or normal (10-18 pint). The default is normal. When you know you will be working with text under 10pt be sure to set the choice to small, otherwise the OCR will not work as accurately as it could. Hint: Be sure to scan small text at 400 dpi. This dpi provides more dots and the scanned characters will therefore be more accurate. Text pitch must also be set. The default is proportional text, or text whose width is determined by its character shape. Fixed text, or monospaced text as it is more commonly known, has characters whose width is identical regardless of character shape. Interactive Learning Interactive Learning is the process of training the program to recognize - Page 43 - new characters. When you wish to use Interactive Learning, select it by clicking on the button to the left. A solid box signifies the choice is active. When Interactive Learning is turned off, Migraph OCR will process images in Automatic mode. In Automatic mode, the program will identify characters as best it can. Characters it cannot recognize, it will replace with either its guess, or the @ sign, if it has no guess. OUTPUT In the Output section you will select the file name and format for saving text and graphics files. Text The Text section determines the file format and style of the text file you'll be saving. Currently the only file format available is ASCII. In the future, additional text formats for specific word processors or DTP programs may be available. NEW - Before processing a file, you need to provide a file name and location where you wish to have the text file saved to. If it is a new file, first select New to designate it as such; then click on the Text Out button. The File Selector will appear. Enter the path and file name that you wish to use. The file extension should be .TXT. Warning: If you forget, and do not specify a file name before processing, Migraph OCR will save the text file to the file named in the Control Panel, overwriting the existing information. APPEND - There will be time when you'll want to add additional text to an - Page 44 - existing file, for instance, when you process a multipage document, or when you've scanned a full page in two or three passes using a hand scanner and want both images in one file. In these cases selecting Append will add the new processed text to the end of the current file. Note: If you know that you will be appending several images, rather than selecting New for the file, select Append. When Migraph OCR goes to save the processed text to the named file and does not find it, Migraph OCR will automatically create it for you. PARAGRAPH - When the Ascii file is saved, this option will put a carriage return at the end of each paragraph. We strongly recommend that you use this option for most of your documents. LINES - When the Ascii file is saved, this option retains the formating (column width) of the scanned text by putting a carriage return after each line. Unless you specifically need this feature, we recommend the Paragraph option since the file it produces is much easier to edit. CACHE (Atari Only) - Page 45 - GET IMAGE, OK AND CANCEL At the bottom of the process panel are three buttons: Get Image, OK, and Cancel. OK confirms changes you have made to the settings. CANCEL deletes changes you have made and restores the original settings. GET IMAGE is the choice you select when you want to bring a new image into Migraph ICR, whether it is importing an existing file or scanning a new one. Depending on whether the Input button is set to Sacnner or File, the Get Image button will react differently. When the chosen Input is File and you click on Get Image, the File Selector will appear. Select the name of the file you wish to load and click OK. The file you selected will load and you will be returned to the Migraph OCR screen. You can now proceed to preview and select the text regions for OCR processing. Scanning directly is handled differently. Migraph OCR does not allocate the memory it needs for scanning until the Scan Area dialog box has been brought up. Even if you do not intend to alter the settings, it will appear the first time in a session that you try to scan. Thereafter, it will not appear unless you bring it up to change the settings. When the chosen Input is Scanner and you click on Get Image, the Scan Area dialog box will appear if you have not yet entered it (through the Settings button). Make any changes you desire and click OK. If you have already entered it, an alert box will appear. It states that the current image will be destroyed and offers you the choice to continue or abort. Select OK to scan, or Cancel to quit. When you select OK, the scanning window light will come on and you may begin to scan your image. Further details on direct scanning may be found in the chapter on Scanning. - Page 46 - MENU COMMANDS This section contains a brief description of all the commands found in the drop down menus. Any command that has an icon associated with it will have the icon pictured next to the command. Keyboard shortcuts are normally displayed in the program's dropdown menus next to the appropriate command, however, because keyboard shortcuts differ between versions, they have been eliminated from the menus shown in this chapter. If a menu entry has (..) or (>>) after it in the drop down, a pop-up menu will appear after you select the command. If the menu entry has three dots (...) after it in the menu, a dialog box will appear after you select it. DESK/MI_OCR MIGRAPH OCR INFO - This command brings up a dialog box with the current version number and copyright statements. PALETTE - (Amiga Only) This command brings up a dialog box that allows you to adjust the screen colors by altering their RGB values. This is useful for reducing screen flicker when you want to run in Interlace mode. Instructions for using the Palette tool can be found under Installation in the Welcome chapter. - Page 47 - FILE ---- LOAD FILE... Use this command when you wish to load an existing image file. The File Selector will appear. Enter the name and click OK. If you have a choic of file formats, we recommend that you load IMG (Atari) or IFF (Amiga) files, since they will load much quicker than TIFF files. SAVE REGION... Use this command to save the currently selected graphic region as an IMG (Atari), IFF (Amiga), or TIFF file. The File Selector will appear. Enter the path and name you wish to save the graphic to and click OK. DEFAULT SETTINGS The following commands all pertain to Migraph OCR defaults. LOAD... Use this command to use previously saved default files. The File Selector will appear. Select the default file you wish to use and click OK. - Page 48 - SAVE... Migraph OCR defaults are saved in the file MI_OCR.DEF that is loaded automatically everytime the program is run. Use the Save Default Settings command when you wish to save the current settings as the default for Migraph OCR startup. SAVE AS... Migraph OCR allows you to save different settings from the Control Panel as default settings. To make a new default file other than the one that will be loaded into Migraph OCR automatically, enter the Control Panel, make any desired changes you wish, and click OK. Then use the Save As command. The File Selector will appear. Enter the path and file name for the default file, using the ile extension .DEF. To retreive any previously saved default files, use the Load command. CURRENT... This command brings up a dialog box which displays all the settings you have selected in the Control Panel. It is an easy way to view your selections. Note that this dialog box is for viewing only, as no settings in it may be altered. The Current Settings dialog box looks like this: - Page 49 - QUIT Use this command when you wish to exit Migraph OCR. SCAN ---- START SCAN This command allows you to begin scanning from within Migraph OCR. If you have not already entered the Scan Area dialog box, it will appear. Make any desired changes and click OK. An alert will then appear; choose OK to turn the scanner on, or choose Cancel to stop. SETTINGS... This command brings up the Scan Area dialog box. Use this dialog box to set the scanning DPI, scan length, unit of measurement, and scan direction (vertical or horizontal). - Page 50 - CLIP ---- The commands in this menu act only on selected clip boxes. UNDO Use the Undo command to "undo" the previous clip command. Because this command acts like a toggle, it can also be used to undo an undo. DELETE This command deletes any selected clip boxes. SELECT ALL This command selects all text and graphics regions. SORT ORDER... Use this command to sort the text regions so they are in correct order before OCR processing. This command is meant for full page scans where more than one or two text regions may be defined. For more information on this dialog box, refer to the Preview chapter. - Page 51 - ERASE GRAPHIC Use this command to erase a selected graphic. # TEXT REGIONS This displays the current number of defined text regions. PROCESS ------- CONTROL PANEL Thsi command will bring up the Control Panel dialog box in which you specify the settings for input, document, output, and cache path. OCR... This command initiates the OCR recognition process on the currently selected text region(s). When iNteractive Learning is selected, the Interactive Learning dialog box will appear on the screen after processing. - Page 52 - ROTATE.. Use this command to rotate the current image 90 degrees to the left or right. Use thsi option when you have scanned an image horizontally outside of Migraph OCR and need to rotate it. INVERT Use thsi command to invert the image. This command is useful when you have scanned white text on a colord or black background. FLIP Use this command to flip the image horizontally or vertically. VIEW ---- Zoom Full This command assumes that the program window frame is set to the full screen, and displays the whole image to fit into that area. Note that if your window is set smaller, you will only see part of the image. - Page 53 - 100% Thsi command shows the image at actual size. As much of the image as possible is displayed in the window. 200% This command displays the image at twice its original size. This is helpful when the image you are working with has a lot of small type. SHOW ORDER Use this command to display the order in which the text regions will be processed. Lines will flow from region to region showing the processing path and a processing order number will appear above the upper left hand corner of the text region. LOCATOR This onscreen icon is used to display the total area of the image in relation to the area of the image displayed on the screen. The white box represents the total area of the image, while the black box represents the portion of the image displayed on the screen. You may use the Locator to view other parts of the image. Simply move the mouse to the black square, depress the left mouse button and move the mouse. The new portion of the image will then be displayed. - Page 54 - INDEX Abort 30 Sort Order 34 Append Dictionary 42 Text 24, 44 Ascii Paragraph format 45 Saving 44 Automatic 44 OCR processing 35 Using 35 Cache 45 Changing Scan Length 41 Characters Defaced 42 Unrecognized 35, 44 Clip Box Combined 32 Commands 51 Creating 28 Defining Irregular 32 Delete 33, 51 Deselecting 31 Irregular 32 Joining 32 Moving 31, 33 Polyline 29 Select all 51 Selecting 31 Undo 51 Close Box 10 Control Panel 39, 51 Defaced Characters 42 Default Cache Path 45 Language 42 Default File 49 Default Settings Get 48 Defaults 41, 48 Save 48 Save settings 49 Delete Clip box 33, 51 Dictionary Append 42 Defaced characters 42 Main 42 New 42 Reading 42 Special 42 Using 42 Document Settings 41 DPI Selecting 41 Drop Down Menus 10 Exit 49 - Page 55 - File Mask 40 File Menu 47 Files Default path 40 Loading 40 Scanning 40 Selecting 40 Full Box 10 Get Image 46 Graphic Region Creating 28 Graphics Defining regions 28 Irregular regions 29 Output 44 Icon Pad 10 Image Viewing 27 Images Defining regions 28 Full Page 30 Invert 53 Loading 48 Loading large 45 Rotate 53 Viewing 53 Viewing 100 percent 54 Input 40 Installation 6 Amiga 6 Atari 8 Interactive Learning 35, 52 Dialog box 36, 43 Using 36 Interafec 9 Invert Image 53 Language Default 42 Selecting 41, 42 Languages Supported 4 Lexicons 4, 36 Lines 45 Load Images 48 Locator 10, 54 Manual How to use 2 Memory Alloctaing 46 Menu Commands 47 Control panel 52 Current settings 49 Delete clip box 51 Get defaults 48 Invert image 53 Load file 48 Migraph OCR Info 47 Number regions 52 OCR processing 52 Rotate image 53 Save defaults 48 Save region 48 Save settings 49 Select all clip boxes 51 - Page 56 - Show Order 34, 54 Sort Order 33 Start Scan 50 Undo 51 Viewing 100 percent 54 Viewing 200 percent 54 Zoom Full 53 Monospaced Text 43 Moving Clip box 33 OCR Automatic 35 Defaults 48 Definition 2 Interactive Learning 43 Main steps 3 Omnifont 3 Preview 27 Processing 3, 35, 36, 52 Processing regions 30 Undrestanding 2 Omnifont Definition 3 Opening Screen 9 Bottom icon pad 10 Close box 10 Drop down menus 10 Full box 10 Locator 10 Slider bars 10 Top icon pad 10 Zoom levels 10 output 44 Palette Adjusting 6 Paragraph Format 45 Paragraph Format 45 Pitch 43 Polyline Clip Box Aborting 30 Creating 29 Practice Session 11 Pretrained Fonts Dictionary 42 Pretrained Typefaces 5 Preview 27 Processing 3 Automatic 44 Interactive 43 Proportional Text 43 Quit 49 Read Dictionary 42 Region Saving 48 Regions Defining 28 Defining full page 30 Defining in order 30 Number 52 Rotate Image 41, 53 Save Region 48 - Page 57 - Scan Area Dialog Box 40, 46 Scanner Contrast 27 Scanning Allocating memory 46 Defaults 41 Direct 40 Horizontally 41 Start 50 Supported Scanners 40 Screen Adjusting the palette 6 Settings 39 Control panel 52 Current 49 Document 41 DPI 41 Hard drive caching 45 Scan length 41 Vertical/horizontal 41 Show Order 34, 54 Slider Bars 10 Sort Order 33 Special Characters Dictionary 42 Starting Migraph OCR From the CLI 6 System Requirements 5 Text Appending 24, 44 Defining regions 28 File format 45 Formats 44 Irregular regions 29 Loading 46 Monospaced 43 Number of regions 52 Output 44 Paragraph format 45 Path 44 Picth 43 Preview 27 Proportional 43 Saving new file 44 Scanning 46 Show Order 34 Size 43 Sorting regions 33 Text Region Creating 28 Text Scanning 46 Training 43 Tutorial 11 Typefaces Dictionary 42 Pretrained 5 Viewing Images 27 Zoom Viewing 27 Zoom Full 53 Zoom levels 10 - Page 58 - That's it!