Explore a groundbreaking real-time sound-based activity recognition system presented at the ACM User Interface Software and Technology Symposium. Delve into the innovative approach of leveraging professional sound effect libraries to enhance acoustic activity recognition capabilities in computing devices. Learn how data augmentation techniques, including amplitude adjustment, reverb application, and sound mixing, exponentially increase the training dataset. Discover the system's performance across various environments and device categories, and understand how it achieves human-level accuracy in real-time activity recognition. Gain insights into the potential applications, limitations, and future prospects of this plug-and-play acoustic technology that could revolutionize context-aware computing.
Ubicoustics - Plug-and-Play Acoustic Activity Recognition
Association for Computing Machinery (ACM) via YouTube
Overview
Syllabus
Introduction
Related work
Why sound effects
How it works
Demo
Results
Limitations
Applications
Questions
Taught by
ACM SIGCHI