Handwriting recognition allows computers to "read" written language.
Handwriting recognition generally refers to techniques used by computers to "read" or translate written text into computer characters. There are two primary types of handwriting recognition: offline recognition, which means the written text has been optically scanned into the computer; and online recognition, which refers to methods where the writing is input with a stylus directly into a computer or touch-screen device linked to a computer (such as PDAs or tablet-style PCs).
Optical Character Recognition (OCR)
Offline handwriting recognition refers to techniques by which a computer scans or "reads" a paper document or text. The common name for such techniques is Optical Character Recognition, also known as OCR. This technique is most successful with printed texts because it is easier to use current character recognition software to recognize printed letter shapes than it is with handwritten material. Although computer scientists and programmers have tried to develop software to read handwriting, this has proven to be extremely difficult.
Challenges of Offline Handwriting Recognition
One of the biggest problems with offline recognition is the lack of standardization among handwriting samples. There can be so many variations from one person's handwriting to another that the computer cannot readily recognize each letter in these samples -- one person's cursive "q" might look like another person's "g," for example. Another major challenge is the fact that cursive letters are linked, and most OCR software is designed to "read" individual characters. It is difficult for such software to extract individual letters from handwriting to identify them.
Online Recognition Techniques
Online techniques usually involve a stylus or pen-like instrument, a LCD screen and software that can translate the movement of the stylus across the touch screen into digital text. One of the most common uses of this technique is in PDAs. PDA users enter information with a small stylus directly onto the screen and the device captures this information digitally in the computer.
How Online Recognition Works
Online recognition programs use complex algorithms to measure, count and otherwise quantify the stylus user's pen strokes into digitized text. Some other versions actually count numbers and arrangements of the pixels on the screen. Such techniques face the same major challenges as the offline techniques, primarily the wide variations in handwriting styles and the problem of distinguishing individual letters from linked cursive words.
The Future of Handwriting Recognition Techniques
Computer companies and academicians continue to work on refining existing techniques and developing new techniques to overcome the problems of handwriting variations. There is no universally successful program for either offline or online handwriting recognition. The greatest success has occurred where the range of characters being "read" is limited; a common example of this is a zip code, where the software only needs to recognize the characters 0-9, the characters are not linked and fewer variations are likely. PDAs remain the most successful implementation of handwriting recognition, but as demand for tablet-style PCs increases, so too does the urgency to develop a consistent method of handwriting recognition.
Tags: into computer, Recognition Techniques, Character Recognition, computers read, Handwriting Recognition Techniques