Speech recognition

Definitions

Speech recognition is the

“

[a]utomated conversion of spoken words into computer-based text. Some speech recognition systems recognize only one person’s voice; others are speaker-independent but recognize a more limited vocabulary. They may recognize continuous speech or, more commonly, require that slight pauses be inserted between words.

”

“

ability to speak naturally and contextually with a computer system in order to execute commands or dictate language.

”

Overview

"Today’s speech-recognition programs build on decades of research in artificial intelligence and speech recognition sponsored by both industry and government. AT&T and IBM Corporation maintained programs in the field, as did DARPA, which seeded early work in hidden Markov models on which the first commercial products were based."

"[T]he technology is far from perfect — the accuracy is diminished by background noise and recording quality, and certain accents are often more easily understood than others."

"The benefits of speech recognition technology are undeniable: hands-free control of technology improves the lives of people with physical disabilities, makes healthcare and other professional services more efficient through accurate voice dictation, enhances automobile safety, and makes everyday tasks more convenient.

"A key feature is that by sending data to the cloud, where powerful computing can be applied, speech recognition services can improve over time. Making use of the huge advancements in data processing in recent years, voice-to-text technologies can now adapt to your speech patterns over time and are getting better at understanding speech in context.

See also

• Automatic speech recognition
• Speech-to-text