OHSU to License Speech Recognition Technology

08/12/03    Portland, Ore.

Seattle area-based Vox Technologies Inc. will market software developed by researchers at OHSU's OGI School of Science & Engineering

Oregon Health & Science University's OGI School of Science & Engineering recently signed a licensing agreement that will enable Vox Technologies Inc. to market a special speech recognition technology called Distributed Voice Processing (DVP). Distributed Voice Processing enables speech recognition in small wireless device (embedded applications such as cell phones) and large vocabulary speech recognition server solutions, such as call centers, voice portals, wireless data services and dictation-based transcription systems.

The embedded speech recognition market is expected to grow from $8.3 million in 2002 to $27.6 million by 2006, according to the Kelsey Group. And according to Bloomberg News, wireless carriers in North America alone will invest $15 billion to entice customers to use cell phones for more than just talking.

The software licensed to Vox was developed at the OGI School of Science & Engineering's highly regarded Center for Spoken Language Understanding. Yonghong Yan, Ph.D., and John-Paul Hosom, Ph.D., both electrical and computer engineering faculty, developed the speech recognition technology.

"It is the goal of our center not just to do good research, but to work on easily transferable, practical solutions that can help the average person in their everyday life," said Jan van Santen, Ph.D., a mathematical psychologist who heads the Center for Spoken Language Understanding. According to van Santen, speech technology could someday be used to help people learn to read, to help nonnative speakers learn English, and to give autistic people more ways to communicate. The center is one of the few in the world focusing on health and education applications of speech technology.

Typical speech recognition systems -- such as dictation or automated airline flight information systems -- translate voice straight into text. "You talk to the computer directly and it understands you," said van Santen.

The software licensed by the OGI school will enable Vox to develop a Distributed Voice Processing technology that has an automatic speech recognition component for small devices, a large vocabulary component for network data centers, and technology that allows the two components to communicate with each other in real time.

For example, existing techniques for dictation require more processing power than is available on mobile devices, such as cell phones. And network service providers cannot offer reliable dictation because of the audio degradation present on normal telephone or cellular voice channels.

"Vox overcomes these barriers by sharing the processing tasks between a wireless device and a server," said Ron Risdon, chief executive officer of Vox Technologies Inc. "No other speech technology company is taking this approach."

Vox heard about the Center for Spoken Language Understanding at the OGI School of Science & Engineering after several key executives attended an educational symposium on campus, said van Santen.

"Vox is interested in our speech recognition system because they need to have access to the source code -- the actual software program -- so they can make modifications for their special DVP technology," said van Santen. "Most licensees with a quality speech recognition system such as ours likely would not be willing to share their source code or would make their binary code exorbitantly expensive. We look forward to a rewarding and ongoing relationship with Vox."

Risdon agreed. "OGI has developed a very elegant and state-of-the-art code base that we think will help us achieve our objective of applying DVP to complex tasks such as navigation services and personal information management systems that are delivered through wireless devices connected to wireless networks."

The Center for Spoken Language Understanding has five full-time faculty, four postdocs, a dozen graduate students and additional programming staff. The OGI School of Science & Engineering (formerly the Oregon Graduate Institute of Science & Technology) became one of four schools of the Oregon Health & Science University in 2001.

OHSU Technology and Research Collaborations (TRC) staff facilitated the CSLU/Vox licensing agreement. Through TRC, OHSU transfers discoveries resulting from its clinical, educational and research activities to companies for commercial development and, when appropriate, creates new ventures. OHSU scientists have disclosed more than 600 new technologies since 1985, resulting in more than $10 million in licensing revenues. This revenue is used to advance OHSU's teaching and research activities.

Vox Technologies, founded in 2002, is based near Seattle in Mill Creek, Wash. The privately held company develops and markets next-generation speech technologies (speech recognition and text-to-speech) that are designed for mobile wireless devices and network data services.