[ad_1]
Handwriting recognition can be done by scanning written text or writing directly to a peripheral input device. Optical character recognition (OCR) is the most effective technique for scanning handwritten documents. Inline recognition has had mixed success, with systems like Newton and Graffiti. Microsoft’s Tablet PCs use a font database. Handwriting recognition research has picked up speed with the integration of pen input in PDAs and mobile phones. Despite challenges, near-perfect recognition is becoming a reality for the mainstream.
Handwriting recognition is most often used to describe a computer’s ability to translate human handwriting into text. This can be done in two ways, by scanning written text or by writing directly to a peripheral input device.
The first of these handwriting recognition techniques, known as optical character recognition (OCR), is the most effective in the mainstream. Most scanning suites offer some form of OCR, allowing users to scan handwritten documents and translate them into basic text documents. OCR is also used by some archivists as a method of converting huge quantities of handwritten historical documents into searchable and easily accessible digital forms.
The second group of handwriting recognition techniques, often referred to as inline recognition, has experienced an ebb and flow in popularity. In the 1990s, Apple Computers released a portable device called the Newton that used the first widely available handwriting recognition interface. Using a small stylus, the user was able to write directly on the Newton’s screen and (in theory) recognize the letters and convert them into text. In practice, the software Newton used to attempt to learn user handwriting patterns was less than ideal, and as a result its popularity was never great.
Later, the Palm company tried a new handwriting recognition system, which they called Graffiti. Rather than relying on an intuitive use of the traditional Roman alphabet, the Graffiti system defined its own system of much simpler strokes of line as substitutes for each letter. This allowed for a higher success rate in identifying letters and learning about a user’s variations, but created a steep learning curve that kept most traditional users at bay.
Microsoft Corporation’s Tablet PCs also use a handwriting recognition system. Rather than attempting to learn a user’s nuances, however, Tablet PCs draw on a large database of font variations. This system appears to have a higher success rate for most users than adaptive systems, but also appears to have a threshold for its reliability.
Research into handwriting recognition software has picked up speed again, with the integration of PDAs and mobile phones with pen input. What was once the realm of fringe tech is fast becoming a multi-billion dollar market, prompting many companies to restart their handwriting recognition investigations.
While there are great problems preventing the creation of a strong and reliable handwriting recognition system, recent discoveries indicate that it is only a matter of time before near-perfect recognition becomes a reality for the mainstream.
