| Sumario: | The article discusses advancements in handwriting recognition technology for tablet personal computers (tablet PC) being developed at Microsoft Corp. The challenges of "pen-based computing" are addressed using a "time-delay neural network" (TDNN) to analyze sloppy handwriting. The recognition process is discussed in detail, with reference to lexicons and the "language-model context." Other topics discussed include personalization, error reporting, addressing the issue of offensive terms, and developing recognition systems in new orthographic language families.
|
|---|