Synaptics Adds to the Handwriting Landscape

Anticipating that there is life in the Windows handwriting recognition market, San Jose, California-based Synaptics, Inc. recently delivered a new print recognition engine that the company claims is over six times more accurate than rival products from Microsoft and Communication Intelligence Corporation (CIC).

The software, dubbed the HR 1200 Handwriting Recognizer, employs proprietary neural network technology. Neural network systems are normally trained through the use of multiple examples, from which the network is typically capable of extracting and encoding the most salient features of the input. In other words, neural-based systems are capable of determining their own rules without the aid of a human expert.

In addition, the nature of neural networks makes the system automatically fuzzy, allowing them to detect important patterns even when there is a less than perfect match. The HR 1200 Recognizer is designed to work with Intel 386/25 MHz systems and higher, though a 486 caliber system is recommended. The software is compatible with Microsoft Windows for Pen Computing and can be swapped in without problem. Synaptics notes that the HR 1200 is designed to be upgradable to Windows 95 when pen-based systems based on this environment become available.

The software is available immediately on an OEM and licensing basis, at a cost of less that US $20 per unit in OEM quantities. Cedar Rapids, Iowa-based Norand will be the first to make the technology available bundled on their PEN*KEY 6600 mobile computing systems and available as an option for the PEN*KEY 6100 and 6300 systems.

Norand cited the HR 1200’s implementation and competitive price point as major reasons for adopting it for its PEN*KEY series. For more information, please contact Robert Sumbs of Synaptics at (408) 434-0110, (408) 434-9819 (fax).