Google MedASR
MedASR is a speech-to-text model based on the Conformer architecture, pre-trained for medical dictation and transcription. It is trained on 5,000 hours of physician dictations in radiology, internal medicine, and family medicine.
UnknownVerified
What is Google MedASR?
MedASR is a foundational model for building healthcare-based voice applications. It accepts mono-channel audio (16kHz, int16 waveform) and generates text transcriptions. It is recommended for dictation tasks involving specialized medical terminologies.
A pre-trained speech-to-text model for medical dictation and transcription.
Key features
- Medical Speech-to-Text
- Specialty Coverage
- Developer-Friendly
- Medical speech-to-text model
- Supports specialized medical terminology
- Foundational Speech Model
- Specialty Training
- Foundational model for voice applications
- Integration with generative models
Use cases
- Medical dictation
- Transcription of doctor-patient exchanges
- Healthcare voice applications
- Medical dictation transcription
- Doctor-patient exchange transcription
- Building voice-based medical apps
- Building custom medical dictation apps
- Specialty-specific transcription development
- Healthcare developers building medical transcription applications
- Hospitals looking to automate medical dictation transcription
- Medical researchers needing to transcribe physician dictations
Who is it for
- Developers building healthcare voice applications