Online handwriting recognition

Abstract

Online handwriting recognition has long been studied, and the technology has already been commercialized to some extent. But limited success stories from the market imply that further research is needed on electronic pen interface in general and recognition methods in particular. This chapter describes some of the basic techniques required to build a complete recognition software. At the turn of the millennium, there has been a renewed interest with focus shifted and diversified to multiple languages. Along with this trend, a lot of new ideas and efforts have been made to deal with different characteristics of different scripts. The difference notwithstanding, it should be helpful for designers to revert to the basics and review the established techniques and new ideas developed from afar the field before figuring out new - or maybe not very new - solutions to one's own language. Current big hurdles in online handwriting recognition include stroke order variation and multiple delayed strokes in addition to shape and style variations. This chapter ends with a short list of example systems and softwares, research based or commercial, that have been sort of landmarks or cited more often than not in the field.