InfraWare Receives Patent for Back-End Speech Recognition
The U.S. Patent and Trademark Office has awarded InfraWare, providers of medical transcription technology, a U.S. patent for the use of phonetic and linguistic features of recorded audio paired with speech recognition results to add new structure and content to speech recognition output.
The patent, #9,377,373 B2, titled "System and Method for Analyzing Verbal Records of Dictation Using Extracted Verbal Features," allows for the inclusion of periods, commas, numbered lists, heading identification, and identifiers for medication lists.
"As a practical matter, this simply means better speech recognition," said InfraWare CEO Nicholas Mahurin in a statement. "Faster turnaround and lower costs are precisely the kinds of solutions every healthcare enterprise needs today."
InfraWare's primary use of this technology is the accelerated creation of healthcare records from physician dictation. The intellectual property involved in this patent forms part of InfraWare's service, known as the First Draft Dictation Recognition Service.
In the InfraWare model, back-end speech recognition is performed to generate a first draft. Medical transcriptionists can then listen to the audio and correct the rough draft as needed.