SFI ADAPT Centre for Digital Media Research

SFI ADAPT Centre for Digital Media Research projects summary

Neuromorphic Vision Sensing for Enhanced HDR Video Acquisition

Primary Researcher:
Paul Kielty

Principal Investigator:
Peter Corcoran

Description:
The research project is centered around event camera’s technology and synthetic events. The project aims to create realistic synthetic event data and developing tools to label event data by using existing networks that are trained on NIR images. The project also aims to improve HDR video and similar technologies using event cameras.

Neural (AI-based) Digital Media Enhancements with Synthetic and Generated Data

Primary Researcher:
Andrei Barchovschi

Principal Investigator:
Peter Corcoran

Description:
The research project is focused on improving digital media speech technologies by increasing the amount of training data for large neural networks focused on speech and language understanding and synthesis. High quality synthetic speech data is essential for training large-scale neural networks in this domain, and generating such data remains a challenge. This project explores various techniques for generating such data using state-of-the-art AI approaches, as well as validating its quality.