Magnum One

home *** CD-ROM | disk | FTP | other *** search

/ Magnum One / Magnum One (Mid-American Digital) (Disc Manufacturing).iso / d13 / pro_cr.exe / MANUAL.DOC < prev next >

Wrap

Text File | 1991-05-15 | 21.2 KB | 509 lines

▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ ░░░░░░░░░░░░░░░ ░░░░░░░░░░░░░░░░ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ PROFESSIONAL OPTICAL CHARACTER RECOGNITION ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ PRO-CR(tm) ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ ░░░░░░░░░░░░░░░ ░░░░░░░░░░░░░░░░ ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ Copyright 1989, 1990, 1991, David P. Gray, Gray Design Associates All Rights Reserved Member, Association of Shareware Professionals ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ C O N T E N T S ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 1. Specification 2. System Requirements 3. Files Distributed 4. Revision History 5. USER GUIDE 5.1 Start-Up Procedure 5.2 Using the Menus 5.3 Top Level Menu 5.4 Options Menu 5.5 Getting a Good Scan 5.6 The Sample Image 5.7 Error Messages 6. Program Self-Check 7. Comments to the Author 8. Association of Shareware Professionals 9. Miscellaneous ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 1. SPECIFICATION ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ * Reads 8 to 30 point mono and proportional fonts * No font selection required * Up to 260 words per minute * Supports HP ScanJet directly * Supports other scanners producing TIFF or PCX files * Selectable resolution, including 200 (fax) and 300 dpi * Preview and online-correction modes with graphics adapter * Real-time viewing of text during processing * Continuous scanning if auto document feeder attached * Mis-recognitions flagged with selectable character * Menu driven or non-interactive mode from DOS command line * Callable from within other programs (Requires License) ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 2. SYSTEM REQUIREMENTS ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ PRO-CR(tm) performs Optical Character Recognition on an IBM PC or compatible. The program will run on an XT or AT, however an AT is strongly recommended due to the highly cpu-intensive nature of the program. A graphics adapter is not required for basic character recognition, but is needed for the preview and online-correction functions. If a graphics adapter is used, it should be CGA, HERCULES, EGA or VGA. (Note for Hercules users: a Microsoft Hercules driver is included. Run MSHERC.COM once before running PRO-CR(tm)). The minimum memory requirement is about 100Kb (512Kb is recommended), although the program adapts itself to use as much conventional memory as available. A temporary disk-file is used for virtual memory for images too large to fit into memory in one go. ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 3. FILES DISTRIBUTED ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ OCR.EXE : The PRO-CR(tm) program README.DOC : Important information HELP1.DOC : Text file used for online help HELP2.DOC : Text file used for online help MANUAL.DOC : This file SAMPLE.TIF : Example TIFF file for processing MSHERC.COM : Hercules driver ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 4. REVISION HISTORY ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 1.0 May 16 1989 : Baseline version. 1.01 May 18 1989 : Fixed character editing in font edit function, caused by bug in compiler's loop optimizer. 1.02 May 31 1989 : Don't reject TIFFs with no bits_per_ sample tag. Assume a value of 1. 1.03 Jun 19 1989 : Don't reject TIFFs with no resolution tags. 1.04 Aug 28 1989 : Fixed bug in learn-mode. 1.05 Nov 29 1989 : Fixed bug in Auto sheet feeder control. 2.00 Apr 30 1990 : Removed font-dependence. Rewrote recognition algorithms. Removed learn and edit functions. Added preview and online correction functions. Added support for 200 dpi, compressed TIFFs and PCX files. Added "unknown" char feature. Installed program self-check. Added DOS command line interface. 2.01 Feb 10 1991 : Removed nuisance delay screen. Fixed problems with command line mode. 2.02 May 14 1991 : Added 1-800 telephone number. ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 5. U S E R G U I D E ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 5.1 START-UP PROCEDURE ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 1. For the interactive Menu driven interface: From the DOS prompt, type: ocr (but see also the options below) 2. For the non-interactive DOS command line interface: From the DOS prompt, type any of the following options: ocr -exec -in=IN -out=OUT -unk=X -dpi=NNN -mode=M Note that all the above options (apart from -exec) are available when starting up in method 1. above. The -exec option is the one that whether PRO-CR(tm) starts up in the menu or non-interactive mode. None of the options are mandatory. IN.................is the name of the image source to be processed. This may be a file name, e.g. image.tif or fred.pcx which is the file name produced by your scanner software. Note that both compressed and uncompressed TIFF files are supported. The TIFF compression methods supported are types 2 and 3 (CCITT group 3 FAX) and 32773 (Macintosh PACKBITS). Alternatively, enter the word SCANNER to process directly from your scanner. For HP ScanJet owners with the automatic document feeder, enter AUTO to process directly using the feeder. The scan will continue until the auto sheet feeder runs out of material to scan. (Note: The keywords SCANNER and AUTO must be entered in upper case). OUT................is the name of the file to write the processed ascii text to. If this file does not exist, it will be created. If this file does exist, text will be appended to it so that you may scan several times to the same file. X..................is the character to use if a mis-recognition occurs. A suitable character would be a "*" or a "~". Not all mis-recognitions can be flagged in this way, there will always be a certain number where PRO-CR(tm) believes it has correctly identified the character. NNN................is the resolution to use when scanning directly from your scanner or when processing from a PCX image file. (The PCX format does not provide useable resolution information). The resolution is entered in units of dots per inch. The range is 100 to 400, but see section 6.4 for help on choosing the correct dpi. M..................is the graphics mode number to specify. PRO-CR(tm) inspects your graphics adapter and selects a suitable graphics mode to operate the preview and online correction modes. In general it chooses the mode with the highest number of colors that has at least 640 pixels horizontally. You may force it to choose another mode by using the -mode option. The IBM graphics modes are: _TEXTBW40 0 /* 40-column text, 16 grey */ _TEXTC40 1 /* 40-column text, 16/8 color */ _TEXTBW80 2 /* 80-column text, 16 grey */ _TEXTC80 3 /* 80-column text, 16/8 color */ _MRES4COLOR 4 /* 320 x 200, 4 color */ _MRESNOCOLOR 5 /* 320 x 200, 4 grey */ _HRESBW 6 /* 640 x 200, BW */ _TEXTMONO 7 /* 80-column text, BW */ _HERCMONO 8 /* 720 x 348, BW for HGC */ _MRES16COLOR 13 /* 320 x 200, 16 color */ _HRES16COLOR 14 /* 640 x 200, 16 color */ _ERESNOCOLOR 15 /* 640 x 350, BW */ _ERESCOLOR 16 /* 640 x 350, 4 or 16 color */ _VRES2COLOR 17 /* 640 x 480, BW */ _VRES16COLOR 18 /* 640 x 480, 16 color */ _MRES256COLOR 19 /* 320 x 200, 256 color */ _ORESCOLOR 64 /* 640 x 400, 1 of 16 colors (Olivetti) */ Note that only modes with 80 column text or at least 640 pixels horizont- ally produce a readable display. Hercules users (mode 8) may need to run the supplied driver MSHERC.COM once before running PRO-CR(tm). Notice: PRO-CR(tm) may NOT be incorporated into any other program (i.e. called from within any other program or batch file) and redistributed without the express written permission from the author. ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 5.2 USING THE MENUS ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ The following paragraphs show how to navigate PRO-CR's menus and read the status line at the bottom of the display. THE MENU BARS......Use the arrow keys to select the required option from the menu bar and hit return when the option is highlighted. Alternatively, hit the highlighted key in the required option to directly select that option. THE STATUS LINE....The line of text on the last line of the screen shows you what parameters you have selected in the options menu and also gives an indication of the percentage completion of the current scan. It looks like this: 0% <ESC=abort> input_file->output_file dpi=200 <PC*> SCANJET <ESC=abort> indicates that you may hit the Escape key during processing to abort the recognition phase. The next two fields "input_file->output_file" indicate both the source of the image to be processed (either from an image file or "SCANNER" if from a direct scan) and the output text file to receive the processed ascii text. Next is the resolution in dots per inch. This must be set if scanning directly or reading from a PCX file. Next is a block of 3 parameters inside angle brackets <PC*>. The "P", if present, indicates that the image "preview" function is active. The "C", if present, indicates that the online correction function is active. The final character is the "unknown character", i.e. the character to output if a mis-recognition occurs. See the help text for the options menu for details of these functions. Finally, the last field indicates the selected scanner type. ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 5.3 TOP LEVEL MENU ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ Top level menu options: Select "Start" to start the recognition process. Select "Auto" to start recognition using the auto sheet feeder. (Note: this option only appears if you are scanning directly and you have the HP ScanJet auto sheet feeder attached) Select "Options" to setup parameters for the recognition process. Options that you need to set up are: - the source of the image to be processed, either a file or the scanner - the resolution of the image in dots per inch (dpi) Note that the dpi may be automatically available if you are processing a TIFF file since the majority of TIFF software includes this. (A warning is given for TIFF files that do not provide this). Otherwise, if you are scanning directly or reading a PCX file, then you need to set the dpi correctly. PRO-CR(tm) has been optimized for the range 200 to 300 dots per inch. ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 5.4 OPTION MENU ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ Select from the following options: INPUT..............Enter the source of the image to be processed. This may be either directly from your scanner (if it is one of the supported types) or an intermediate image file produced by your scanner software. To select direct scanning, enter the word "SCANNER" here. To select a file, enter the name of the image file produced by your scanner software. Currently supported types are TIFF (both compressed and uncompressed) and PCX (PC Paintbrush) files. Be sure to include the extension in the file name (i.e. .TIF or .PCX) so that PRO-CR(tm) can determine the file type. For the case of PCX files, it is necessary for you to set the resolution that the image was scanned at (since the majority of PCX producing software does not include this in the file). Use the "dpi" option to set this. OUTPUT.............Enter the name of the file you wish PRO-CR(tm) to write the ascii text into. Note that PRO-CR(tm) will create the specified file if it does not already exist, and will append text to the specified file if it does exist. DPI................Enter the resolution in dots per inch. There are 2 cases when this is important. (1) When you are scanning directly. This tells the scanner what resolution to scan at. (2) When you are processing an image file and the file does not contain the image resolution. This is always true of PCX files and is sometimes true of naughty TIFF files. (It is rare for TIFF software to omit the resolution and you may assume it is present unless PRO-CR(tm) warns you otherwise. Never assume a PCX file contains the image resolution though !). Knowing how to set the right dpi for DIRECT SCANNING: PRO-CR(tm) has been written to run best at 200 to 300 dpi. 200 dpi covers the fax standard and most hand scanners. (Some popular hand scanners claim 300 and 400 dpi modes, however many simply duplicate pixels to match printer resolutions and therefore add no more information to the scan for OCR purposes). For reasonably sized text 200 dpi should be adequate, but if the text is tightly spaced (runs together) then 300 dpi should help. PRO-CR(tm) allows a range of 100 to 400 dpi for experimentation purposes. In general, try 200 dpi first and then 300. UNKNOWN CHAR.......Enter the character you wish to be embedded in the output text to flag characters that could not be recognized. This is useful if you use an editor after the recognition phase to correct mis-recognitions. It allows you to search for this character to speed up the correction phase. Use of a spelling checker should also prove helpful. PREVIEW............This function is active if a check mark appears next to it in the menu. The function is toggled on and off by repeatedly selecting it. When active, as much of the image as can fit on the screen is displayed prior to being processed. This is useful to examine the quality of the scan, for skewed paper, lack of contrast, suitable resolution etc. This function is only available for users with a suitable graphics adapter. CORRECT...........This function is active if a check mark appears next to it in the menu. The function is toggled on and off by repeatedly selecting it. When active, you will be allowed to correct mis-recognitions online during the OCR processing phase. You will be prompted for the correct character(s) when a mis-recognition occurs. At the lower right hand side of the screen an image of the mis-recognized character(s) will appear for you to correct. If you are unable to recognize the text, hit return and the text will be replaced by the "unknown" character mentioned above. You may also hit Escape to abort the processing. This function is only available to users with a suitable graphics adapter. SCANNER TYPE......Select your scanner type from the list. At the time of writing, only the HP ScanJet is directly supported, however I have had a large number of requests for the Logitech SCANMAN and plan to support this. ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 5.5 GETTING A GOOD SCAN ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ To get the best scan, several factors must be taken into account. 1. Make sure the text is scanned in a straight line. This is especially crucial for users of hand scanners. The lines of text should appear as horizontal and level as possible when previewed. If lines are skewed, or rotated, not only is the recognition accuracy impaired but it may cause PRO-CR(tm) to mistakenly run lines together. This is not so much a problem for flat bed scanners since they have guides for inserting the paper. Even so, some printers may print sloping lines which are not perpendicular to the sides of the paper. Hand scanners have the additional problem of matching up "strips" when scanning in several passes. I would suggest that 1 strip at a time be processed in this case. 2. Make sure the scanner's contrast is set to a suitable value. This is not a problem for the HP ScanJet which has an auto-contrast feature. With other scanners, preview the image and look for broken or faded characters or extraneous "noise" as a result of poor contrast. Best results are obtained when the background is clear and the characters appear sharp. 3. Select a suitable dpi. Use preview to inspect the image. If possible, select a dpi which causes the characters to appear separate. If too many characters run together, the recognition accuracy will be impaired. If you cannot separate the characters using a higher dpi (inspection of the original copy will show that the characters are indeed joined together) then select the online-correction mode and you will be prompted to enter the unrecognized portion of each scanned line during the recognition phase. If you do not select the online correction mode then PRO-CR(tm) will attempt to separate characters that are joined together up to a maximum of 3. For mono-spaced fonts this strategy works pretty well, but for proportional fonts results are less desirable. Note that for some near letter quality dot matrix fonts, lowering the dpi will actually improve results. The reason is that the dots making up the characters will appear to join up and become united. PRO-CR(tm) cannot read draft dot matrix fonts since the dots making up the characters are disjointed. For the same reason, PRO-CR(tm) will not read characters that appear broken. Broken characters lose their essential features which PRO-CR(tm) relies upon for recognition. ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 5.6 THE SAMPLE IMAGE ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ A sample image is supplied with which you can experiment with PRO-CR's features. The image is in the file SAMPLE.TIF and is in compressed TIFF format. Make sure the status line shows sample.tif as the image source. Select the online correction feature and process the image. Notice that the "mm" in the word "common" and "qu" from the word "quality" are presented for correction. This is because these characters are joined together and cause a mis-recognition. After the scan, deselect the online correction feature and reprocess the image. Notice that PRO-CR(tm) successfully separates the "mm" by itself but fails with the "qu", only recognizing the "q" correctly. ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ 5.7 ERROR MESSAGES ▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒▒ The following error codes may be seen, to do with TIFF files. 1 ...... Could not find the input file. 2 ...... Non-Intel byte order. The TIFF file is possibly a Mac file. 3 ...... Wrong value for bits_per_sample tag. 4 ...... Unsupported Compressed TIFF file. 5 ...... Wrong value for photometric_interpretation tag. 6 ...... Wrong value for fill_order tag. 7 ...... Wrong picture orientation. 8 ...... Wrong value for samples_per_pixel tag. 9 ...... Wrong value for minimum_sample tag. 10 ...... Wrong value for maximum_sample tag. 11 ...... Wrong value for planar_configuration tag. 12 ...... Missing bits_per_sample tag. 13 ...... Missing image_width tag. 14 ...... Missing image_length tag. 15 ...... Missing image_pointer tag. 16 ...... Missing X_resolution tag. 17 ...... Missing Y_resolution tag. In addition, there are several self explanatory warnings and other error messages. ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 6. PROGRAM SELF-CHECK ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ PRO-CR(tm) provides a measure of protection for itself against accidental corruption during downloads from bulletin boards. A self-check is performed when the program starts up and will notify the user upon failure. Note that the program must be run from the directory where it resides or the self-check will fail. ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 7. COMMENTS TO THE AUTHOR ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ Any feedback would be greatly appreciated. Please direct any comments to the author personally via mail to David P. Gray, Gray Design Associates, P.O. Box 333, Northboro, MA 01532, USA. ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 8. ASSOCIATION OF SHAREWARE PROFESSIONALS ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ This software is produced by David P. Gray who is a member of the Association of Shareware Professionals (ASP). ASP wants to make sure that the shareware principle works for you. If you are unable to resolve a shareware-related problem with an ASP member by contacting the member directly, ASP may be able to help. The ASP Ombudsman can help you resolve a dispute or problem with an ASP member, but does not provide technical support for members' products. Please write to the ASP Ombudsman at 545 Grover Road, Muskegon, MI 49442, USA or send a CompuServe message via Easyplex to ASP Ombudsman 70007,3536. ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 9. MISCELLANEOUS ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ HP and ScanJet are registered trade marks of Hewlett Packard. ScanMan is a registered trade mark of Logitech inc. ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ END OF MANUAL.DOC ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓