Content area
Full Text
Virtually everyone who has used a flatbed scanner has been desensitized to the magic of optical character recognition (OCR) technology. If there's a page of text from a book or journal that you want in electronic form so it can be edited, you simply place the page on the scanner and the OCR software does the rest. The software converts the picture of each character, which is composed of a predictable black and white pattern, into its keyboard equivalent.
The primary issues with OCR are accuracy, speed, and formatting. Accuracy is a function of both the software and the quality of the scanned image: The better the image quality, the greater the accuracy of recognition. Although even the best OCR software occasionally confuses a "c" with an "e," especially if the original is smudged, accuracy typically is near 99%.
Speed, in pages per minute, is primarily a function of the scanner speed but can also be affected by PC processor speed. Most affordable scanners can process a page in about 15 seconds. The OCR process can take as little as 5 seconds on a modern PC running at 500 MHz or up to 30 seconds per page on an older 200- MHz PC. Formatting or retaining the formatting of the original document is strictly a function of the OCR software. Even the least expensive OCR software can recognize the typical multicolumn journal article.
With the price of quality flatbed scanners hovering around $250, which usually includes a limited version of OCR software, typing in documents simply isn't cost-effective anymore. If you're primarily scanning an abstract occasionally or even a full article, then the limited OCR software that comes bundled with many flatbed scanners should be more than adequate. However, if you make a habit of scanning tables, spreadsheets, and other complex documents, then upgrading to a full version of an OCR package, such as OmniPage Pro (about $90), makes sense.
Behold the power of the pen
Although it still has a...