Scanning is the process via which documents and pictures on paper (e.g., an old book or a hanging photograph), for which no digital version (nowadays also called 'the soft copy') is available, need to proceed so as to remain in your computer or to be available through the Internet (or, may be, through compact disks). A scanner converts the paper document generally to digital image files, though direct one-step conversion to digital text by simultaneous use of OCR technology is also sometimes performed. However, the two-step conversion process (first the image files and then the text documents) is generally preferred for on-paper text matters, as this can make both the image files and the OCR-output text available to the operators - they may then compare the two to check the correctness of the text and to manually introduce some corrections to the text as is found necessary. This is how things are being done by the volunteers working for the great book-digitisation initiatives such as Project Gutenberg: one scans some pages of a book and deposit the image files to the online repository, another one performs the OCR conversion to get the text files and then deposits that to the online repository, while a third one downloads both sets of files and compares them to manually correct any errors that has crept into the text during the OCR conversion process, and then deposits the corrected text. [If you want to join the said worldwide group of volunteers, click here.]
Note: OCR (Optical Character Recognition) is the modern-day computer based technology with which the digital images of textual characters (i.e., of the alphabets, digits and symbols such as +) can be recognized as those characters themselves. It is clearly a great advancement of computer technology to be able to recognize the characters within an image and obviously requires some degree of artificial intelligence. Free, effective and easily accessible software for performing OCR have started appearing only as recently as in the new millennium (we're going to discuss some of them)! But yet there may occur a few errors here and there while performing automated OCR, so a final finishing touch of manual correction on such OCR-output still remains necessary.

        Using a scanner, scanning may be done using the proprietary software of the scanner producer (e.g., HP, Inc.) distributed along with the scanner, but the operating system (e.g., Windows or Fedora Linux) generally provides its common scanning program. Thus Windows OS has Microsoft Scanner and Camera Wizard, whereas Fedora has Simple Scan. Opening that sort of a program, you may choose two basic types of scanning options as: (a) color (colour)/ grayscale (greyscale)/ black & white scanning and (b) fine or low resolution (implying high or low quality, and thus also larger or smaller file size) - it is measured in dpi ('dots per inch') unit, with 300 dpi (also called 300 x 300 dpi) meaning fine, 200 dpi the usual and 100 dpi rough quality. For a coloured photograph, you need to obviously go for a colour scan option - otherwise you'll arrive at a monochrome photograph image! But for scanning a text document with a final aim to go for its OCR, even if the document was printed in blue ink, you may safely go for the greyscale scan (thus greatly reducing the image file size) - knowing that colour or greyscale scan would hardly be of any difference as far as the final OCR result is concerned. But, remember, the black & white option (found in case of Windows), in contrast to the greyscale option is generally disastrous: I found practically nothing (while seen with the eyes) after thus scanning the first page (A4 sized) of a novel freshly printed on white paper and also nothing meaningful as its OCR output, whereas a corresponding greyscale scan gave a rather good image and an OCR output with only seven errors (illustrated later)!


        About scanning, I've observed something really curious and useful as well, which I should share with you now. Using my HP Scanjet 2300c scanner (a common one, besides what I'm going to disclose is most probably true with other scanners also), I uniformly found that the scanned page images obtained in my Fedora is much clearer as well as pleasantly readable compared to those obtained in my Windows (see below). It's not something about the dpi resolution option (I've carefully tested with 300 dpi in both cases), nor it has something to do with the colour/greyscale option (I used the colour option in both cases). Just see the unbelievable difference in the images below (they were reduced to 25%, so as to be viewable within this window), corresponding to a recipe offered by a local Guwahati bakery. The large-sized scanned images actually obtained are being stored here and here for your reference (to save any of these images to your computer for practicing OCR, click to open the desired webpage and then right-click at the large image, followed by clicking at the Save Image As... option). And, did it take a significantly longer time to scan in Fedora? Though I haven't tested that with a stopwatch, I didn't feel anything like that - rather felt that Fedora took a somewhat lesser time!