Introduction:-

Welcome to this study,

Our goal is to train an Automatic Speech Recognition (ASR) model using synthetic medical conversations between patient and physicians. This would be an in-person recording to be done in a clinical set up. 

Note Project Limits, and Demands:

• Recordings would be done in Shaip APP.

• Scenarios for recording would be provided by Shaip which covers family adult medicine specialty

• One can provide up to 4 hours of unique synthetic dictations

• Recommend planning to help you reach that goal

• Example: 10 minute conversation x 6, repeat each day for 4  days

• Approach for recording:

• Physician has to download the APP  Google Play Store or Apple App Store,

• You need to use the provided the user name and password for the login once you download the App. 

• Patient has to review the prompt

• Keep the phone on a table at a equidistance from patient and  physician

To be successful you must meet these quality requirements: 

• Adherence to Prompt:

• Prompts/Scenarios would be displayed in App

• Person playing a patient role has to understand the scenario in advance before starting the recording

• The patient has to portray the character as per the prompt/scenario mentioned in app while recording

• Include pseudo PHI (fake Protected Health Information)

• Please use pseudo PHI, as mentioned in the prompt

• Quality

• Clinical set up

• Conversation should be spontaneous and complete

• Clear and accurate pronunciation

• Adequate speech volume for both the participants