Outcomes

Demos

Playlist on our YouTube page
It includes demos of systems we made:

• Automatic Speech Recognition: audio to text with confidence-based editing interface (2024)
• Spoken Term Detection: speak a word and this system will find whereever that word is spoken by anyone in a video database (2024)
• Audio Fingerprinting: give a short audio recording and this system will find the original audio from an audio/video database (2024)
• Text-based Event Search: write a word (e.g., horror music, clapping, lion) to find audio/video containing that kind of audio from an audio/video database (2024)
• Live Sound Event Detection (2023)
• Music Education: Smart platform for teaching music that automatically detects singing mistakes in a learner by matching with teacher’s singing (2023)
• Voice Assistant for Child-birth: ASR based interaction with nurses, delivery room management and tracking of mother’s vitals during delivery (2022)

Datasets

• MADHAV Lab Mistake Detection for Music Teaching (M3) Database

• MADHAV Lab Tic Database

• Ornamentation detection dataset (ROD):
  • Dataset link
  • Paper link
• HAR: Hindustani Alankaar and Raga Dataset:
  • Dataset link
  • Paper link

• M3 dataset: Not public yet

• Hindi ASR dataset:
  • Dataset link
  • Paper link

Softwares

• AdvNF for sampling Boltzmann distributions
• ScoreNF for sampling Boltzmann distributions
• Automatic Music Tutor

For other softwares and codebases to our projects, please visit our github page.