Computational and Media Engineering

Work in digital media applications includes spatial audio for VR, speech analysis and understanding, and visual algorithms for cinema post-production. Fundamental work in Bayesian inference underpins much of what we do.

The groups involved have had fantastic success in the last five years in embedding technologies into mass-market platforms owned by Google and YouTube. In addition, our visual algorithms research is used by every post-production house working on movies today.

The Signal Processing for Media Applications Group (Sigmedia) was founded in 1998 when Prof Kokaram joined the Department. Since then it has expanded to include Prof Naomi Harte, Dr François Pitié, Dr. Enda Bates, and Frank Boland (Emeritus Prof). The group's expertise is in media signal processing including motion picture FX, speech and audio processing, and audio-video signal analysis. Group members have won numerous industry and academic awards including an Oscar.

The course in MMT provides a symbiotic link with the music creation sector.

This sector hosts about 20 Ph.D. students and 5 postdocs. We also participate in EU research projects and SFI Research centres. Together with strong industrial links, this income has averaged about 1M€ per year.

Principal Investigators:

• Prof. A. C. Kokaram - image processing, video restoration, digital television.
• Dr. A. P. Quinn - speech recognition, speech synthesis, Bayesian techniques.
• Dr. N. Harte - video processing, speech recognition, digital systems design.
• Dr. F. Pitié - signal media processing.
• Dr. H Javidnia - depth perception and 3D reconstruction, visual inertial odometry, event-based vision and navigation, conditional generative models for synthetic data generation.