Speech Technology Magazine


NovoSpeech Scores Investment

Funding will be used to ramp up business development.
Posted Jan 3, 2013
Page1 of 1
Bookmark and Share

Israeli-based NovoSpeech has received close to $500,000 in funding, which it will use to continue development of its flagship product, Speech++, as well as to ramp up sales and business development efforts for its Pulsar1000 and Quasar1000 products.

The funding comes from the Maryland/Trendlines Israel Fund, the Mofet Venture Accelerator, and private U.S. investors, as well as the employees of the company.

NovoSpeech analyzes the bottlenecks inherent to existing speech recognition solutions and has developed a technology that improves existing algorithms and imports algorithms from other disciplines, resulting in a proprietary, robust, speaker-independent speech recognition technology with superior accuracy rates.

Among NovoSpeech's features are:

Highly noise-robust recognition accuracy, including high accuracy even in real-life (noisy) environments. Benchmarking results show a 35 percent improvement in NovoSpeech's accuracy as compared to state-of-the-art ASR.

Dramatic reduction of computational complexity. In addition to the ability to deploy on low-end architectures (such as Bluetooth BC05), NovoSpeech will be able to deploy large-vocabulary continuous speech fully device-side on mobile devices, eliminating the constraint of current ASR mobile solutions that must be connected to the cloud.

Accurate identification of words regardless of gender, accent, or speaking style.

Page1 of 1