The document is loaded into the scanner with the text down, towards the scanner glass. The general rule is this. The top of the text should be where the scanner lamp starts from. This will be the correct portrait orientation. In portrait orientation, the scanner lamp passes text from top to bottom. And the correct landscape orientation of the text is obtained by rotating 90 degrees clockwise. That is, in landscape orientation, the scanner lamp should pass through the text from right to left or against the reading direction. If the sheet is laid correctly, then you can turn off automatic page orientation detection, which will speed up text recognition.
 Scanning and recognition.
FineReader contains in its menu various commands for scanning and OCR. The simplest of the commands is “Scan and Recognize” or simply “Scan”. This command will scan and immediately recognize one page of text. After the recognition is completed, the newly recognized text will appear in the Text window. The cursor will be placed at the beginning of the recognized text. Jozu usually responds to this by saying the word “Text”. After that, the recognized text can be read and edited as in a regular text editor.
If you are scanning an entire book, it is convenient to scan all the pages first and then recognize them.
Let’s list the main commands for scanning and recognition (for FineReader 6.0) (ABBY likes to greatly change the menu in different versions of FineReader, as well as the structure of the options dialog, so hot keys for basic operations may differ in new versions of FineReader and be located in other menu items). In version 11, the scanning process generally takes place in a separate dialog box, which is opened by the “Scan Pages” command from the “File” menu.
“Control + d” – scan and recognize one page or book spread (Process menu)
«Control+k» – scan one page (File menu).
“Control+Shift+k” – scan many pages (menu “File”).
“Control+T” – stop scanning (“File” menu).
“Control+r” – recognize one page (menu “Process”).
“Control + Shift + r” – recognize all pages (menu “Process”).
“Scan and recognize multiple pages” is a command from the “Process” menu.
“Control+Shift+s” – calls the Scanner Settings dialog (in the “Tools” menu, the “Scanner Settings” item).
If you are scanning a book spread (landscape text orientation), you can set the Divide Book Spread option on the Scan tab of the Options dialog. This will cause FineReader to split the two spread sheets into two different files after scanning and recognize them separately. This method will improve the quality of recognition and reduce the number of errors. You can disable this option when scanning individual pages.
However, FineReader may not understand that there is a spread of a book in front of it and leave this spread as a page of text in two columns.
And it happens even worse. If the text consists of many columns, FineReader may divide the spread not in the middle, but in another place. And then you will get half a page and one and a half pages of text and incorrect recognition is guaranteed. Therefore, be careful with the “Divide Book Spread” switch. Apply it only to quality text in a well-folding book.
At the place of the spread, the paper usually bends. The place of the bend after scanning looks like a dark stripe. If letters fall into the bend, then in this place the image of the line is bent, and the letters themselves are distorted. Still letters can fall, as it were, into the gray shadow of the bend, partially or completely lost. Numerous errors at the ends or at the beginning of lines indicate that the book needs to be pressed harder on the spread to reduce curl.
If the book does not unfold well enough, switch to gray scan mode. This will reduce the influence of the bend shadow on the outermost letters and improve recognition.
When scanning a spread of a book in the “Scanner Settings” dialog, you must set “Landscape Image Orientation”, and when scanning one normal page of text – “Portrait Orientation”. True, new versions of Fine Reader already recognize text orientation well in automatic mode.
Text of normal size and print quality can be scanned at Manual Brightness 50%. If the letters are small, increase the Resolution to 400, 500 or 600.
If the print quality is poor or the book does not unfold at all, go to “Auto Brightness” and set the “Picture Scan Mode” to “Gray”. This will slow down the scan, but improve the recognition quality.
All these parameters are set in the “Scanner Settings” dialog (“Control+Shift+s”).
If there are gray spots on the sheet, or letters and graphics on the back of the page are showing through due to thin paper and bold black text, increase the scan brightness to, for example, 60 or 70. This will make the spots and letters show through from the back of the page disappear and drawings.
If the letters are too faint, then lower the scan brightness, for example, to 40. This will cause the letters to become clearer and bolder, and they will be easier to recognize.
If the text on the sheet is arranged in several columns, as in a magazine or newspaper, then on the “Scanning” tab of the “Options” dialog, the “Page Type” switch must be set to “Auto”. And if this is a regular book text, then it is better to set the “Text Type” switch to “One Column”, which will speed up recognition. To be sure, it’s best to always keep this switch in “Auto” mode.
Comment. If the “Divide Book Spread” checkbox is set, then the book can be recognized by setting the “Text Type” switch to “One Column”. If the “Divide book spread” flag is not set, then both pages of the book are recognized as one sheet of text. That is, it turns out the text typed in two columns. Therefore, in this case, you must set the “Text Type” switch to “Auto” so that the text is recognized as consisting of several columns.
If you need to rescan some pages, then on the “Scanning” tab of the “Options” dialog, check the box “Prompt for a page number before adding to a batch.” This will enable you to match the scanned page number in the FineReader package with the page number in the book. Then this option can be canceled, since further pages will be numbered in order and added to the end of the package.
If the text is recognized very poorly or the recognized text is completely unreadable, first check the page orientation, that is, whether the page is laid correctly in the scanner. Next, check the very first parameter “Recognition language” on the “Recognition” tab of the “Options” dialog. An incorrectly chosen language can lead to a huge number of errors. Do not set the “Russian-English” recognition language if there are no English letters in the text, so that Russian letters are not replaced with similar English letters. And only after checking the recognition language, increase the scanning resolution and switch to the gray scanning mode.