The position of the pen tip can be measured accurately with current tablet technology (Teulings & Maarse, 1984; Meeks & Kuklinski, 1990). This makes it a suitable signal for scientific research on the handwriting process, as well as providing a useful signal for handwriting recognition. The main advantage of on-line handwriting recognition over off-line (optically scanned) handwriting recognition is that the stroking order is known, and spatially overlapping letter parts are well separated in time. Also, the results of handwriting recognition can be seen immediately and corrections can be made during the interaction. However, the dotting of the i and j and the crossing of a t can be a problem, because they are postponed in time, e.g., until the end of a sentence. To the user of pen computers, the time axis is considered irrelant, only the visible ink counts, whereas this may not be the case for the underlying recognition algorithm. Another disadvantage is that dedicated equipment (digitizer or pen computer) must be present at the time of handwriting production. In off-line handwriting recognition, the special equipment is needed at the receiving end, at the time of scanning. As a consequence, much more scanned handwriting data are available for the developement of optical character recognizers than is the case for on-line handwriting recognizers: there is a vast amount of handwritten material on paper which can be used. Still, on-line handwriting has a number of attractive features. It allows for small keyboardless computers, and entry of all essential data types (commands, text, graphics) in an interactive way. Too counteract the problem of "data starvation" (Yann Le Cun, personal communication), an international project was founded, with the goal of collecting on-line handwriting data (UNIPEN).
The goal of this page is to show a unique feature of on-line handwriting recordings: the calculation of pen-tip velocity. This time function reveals a human writer in action. It will take some time to load the applet and the data so please be patient...
Please refer to our Publications when using anything from the shown material.
to NICI's Java-based Handwriting Demos page
to NICI's on-line handwriting recognizer description
Copyright Lambert Schomaker (April 1, 1996)