IMVIP 2023

                       Petronel Bigioi

Driver Monitoring Systems and Next-Generation Perceptual AI

Petronel Bigioi is Xperi's CTO of Product, responsible for leading the engineering team focusedon developing image processing and audio solutions for the home, automotive and mobile markets.

Petronel has more than 250 granted and published U.S. and international patents to date.Petronel is an IEEE fellow with over 20 years of experience in the digital still camera and mobile phone industries, working in both signal processing and connectivity. A co-founder of several successful companies (including FotoNation), he is also a pioneer of digital camera connectivity, a co-author and actively involved in the picture transfer protocol (PTP) and PTP-over-IP networks communication standards. 

Petronel got his Bachelors of Science in electronics engineering (1997) from Transilvania University, Brasov. Petronel also holds two master degrees in ASIC design (Transilvania University,  Brasov, 1998) and Networks and Communications (National University of Ireland, Galway, 2000). Petronel obtained his Ph.D. in electronics from Transilvania University (2005), his work being recognized by Romanian academy of science with a “Gheorghe Cartianu” award.

Fake Children: Why we need them and how to make them.

Researchers working with human-centric Machine Vision applications have found GDPR tobe a huge challenge. Many of today's Machine Vision systems rely on neural network modelsand require large training datasets for optimal performance. But what do you do when your
application is for a Smart-Toy and you need data from a vulnerable population, such as youngchildren, in order to train your Edge-AI Machine Vision system? GDPR imposes manycomplexities in collecting, managing and processing data from real children.

Fortunately it is now quite feasible to leverage state-of-the-art GAN and other generative neuraltechnologies to build data samples at scale. In this talk we'll get some insights into the powerand potential of today's neural technologies to build a gender balanced dataset of child data, including controllable facial expressions, age variations, facial pose and even speech-drivenanimations with photo-realistic lip-synch. Learn more from our second keynote presented by Prof. Peter Corcoran and Muhammad Ali Farooq from the University of Galway.

P‌rof. Peter Corcoran