Note Buddy (in Collaboration with SingHealth)

Overview

Improved the Note Buddy speech-to-text systen with real-time transcription, speaker diarization with text summarization using OpenAI GPT-4o model and Azure Speech SDK.

Links :

NoteBuddy Final Presentation : slides

Code on GitHub : code

News : SingHealth news

Topics

1. Real-time speech to text transcription
2. Speaker diarization
3. Text summarization
4. Evaluation of speech models and performance benchmarking

Outcome & Impact

Results for this semester were released this morning and I am thrilled to share that our team (Team 9) secured an A+ for our Capstone project 🎉

We partnered with SingHealth to improve Note Buddy - a digital scribe designed to transcribe and summarise clinical consultations into medical notes in real time. Our goal was to reduce the doctors’ administrative workload and enable them to focus more on patient interactions and care.

This achievement would not have been possible without the guidance and expertise from our industry advisors at SingHealth. A heartfelt thank you to Mark Heng, Eileen Lew, Ajeet kumar for your unwavering support!

Finally, a big shoutout to my incredible teammates - Faith Tan LH, Zandra Tay, Adelle Low and Pearlyn Liu. Their dedication, analytical mindset and technical skills were pivotal in tackling the challenges we faced over the past four months, ultimately driving the success of this Capstone project.