border border

QuickScribe: An End-to-End Transcription Tool

Team members

Tan Yong Da Dino (ESD), Matthew Lau Zhi Yuan (ESD), Leong Yu Fong (ESD), Ong Sze Qi (ESD), Teng Jun Yuan (ISTD), Ng Jo-shen (ISTD), Cheow Pak Leng Josiah (ISTD)

Instructors:

Georgios Piliouras, Liu Jun, Natarajan Karthik Balakrishnan

Writing Instructors:

Rashmi Kumar

Teaching Assistant:

Siddharth Kumar

mockup

PROBLEM OVERVIEW

Speech-To-Text involves the conversion of an audio file into a text transcription. This is a tedious and time consuming task. Home Team Science and Technology Agency is currently adopting a fragmented approach in Speech-To-Text through the use of various audio editing and transcription softwares, whereby the main transcription process is outsourced. This leads to high upfront costs and man-hours incurred.




OUR SOLUTION

QuickScribe significantly improves and expedites the transcription experience. It integrates Audio Processing, Automatic Speech Recognition, transcription toolpage, and text analysis into an end-to-end pipeline.


key functions heading


“1”

“2”

“3”

“4”


11
transcription result
12
spech recognision process



16
interface
database
analytics
states
states
htx
FEEDBACK

TEAM MEMBERS

student Tan Yong Da Dino Engineering Systems and Design
student Matthew Lau Zhi Yuan Engineering Systems and Design
student Leong Yu Fong Engineering Systems and Design
student Ong Sze Qi Engineering Systems and Design
student Teng Jun Yuan Information Systems Technology and Design
student Ng Jo-shen Information Systems Technology and Design
student Cheow Pak Leng Josiah Information Systems Technology and Design

© 2021 SUTD

border border