Background
Automatic speech recognition systems (ASR) are designed to accurately convert spoken words into text. As a therapeutic application, automatic speech recognition (ASR) has the potential to assist speech pathologists and their clients by providing feedback on articulation consistency.
However, there are certain limitations to commonly available ASR systems. To be used as feedback for speech and language therapy, these systems must covert the spoken words “as is”, even when articulated incorrectly so as to provide appropriate feedback. This requires highly sensitive systems in order to recognize subtle deviation in articulation.
In addition, many ASR systems are domain specific and can not identify large vocabularies; they are also speaker dependent and require cumbersome training in order that the system will recognize the client’s voice.
To date, there is no speech-language therapy software suitable for the mass market that fully utilizes the capabilities of automatic speech recognition.
Solution
NovoSpeech has developed a natural language speech recognition system based on a unique encoding engine that combines a new approach to language understanding and acoustical analysis. The result: a robust system that offers significant improvement in the phoneme recognition rate and in the ability to accurately identify articulation for various speech disorders.
Value Proposition
NovoSpeech’ has developed an engine that is the foundation for many therapeutic applications in addition to speech therapy such as assessment of speech disorders and training deaf people.
In addition, the Company’s solution brings high quality speech therapy to individuals who live in remote and rural areas or who can not afford speech therapy lessons.
NovoSpeech’s proprietary natural language speech recognition engine—
- accurately identifies — in real time —all words spoken by any person, independent of noise, accents, and vocabulary size
- recognizes subtle differences between articulated word and text and provides precise feedback
- suits the mass market; is speaker independent
Target Market and Approach
NovoSpeech has made a strategic decision to initially target the speech-language therapy market, which includes: speech sound disorders, such as articulation, stuttering and aphasia, and language disorders as well as dyslexia conditions.
NovoSpeech will partner with speech and language therapy distributors and speech therapists and hospitals.
U.S. Market size
According to the U.S. National Library of Medicine¬
- by first grade, about 5 percent of children have noticeable speech disorders
- more than three million people stutter
- about one million people suffer from aphasia
- between 6 and 8 million individuals have some form of language impairment.
- Speech-language pathologists held about 110,000 jobs in 2006.
|
ICT - Information and Communication Technology
