Antonino Furnari is a research fellow at the Department of Mathematics and Computer Science of the University of Catania and member of the FPV@IPLAB research group. He received his PhD in Mathematics and Computer Science in 2017 from the University of Catania and authored two patents and more than 50 papers in international book chapters, journals and conference proceedings. Antonino Furnari is involved in the organization of different international events, such as the Assistive Computer Vision and Robotics (ACVR) workshop series (since 2016), the International Computer Vision Summer School (ICVSS) (since 2017), and the Egocentric Perception Interaction and Computing (EPIC) workshop series (since 2018). Since 2016 he has held talks at universities, including the University of Bristol (UK), the University of Bern (CH), the University of Essex (UK), ETH (CH), at workshops (EgoApp 2019 in conjunction with BMVC 2019), and tutorials at conferences (VISAPP 2019, VISAPP 2020, VISAPP 2021). In 2020, he has been guest editor for IEEE Transactions on Pattern Analysis and Image Intelligence (TPAMI) with a special issue on “Egocentric Perception”, and in 2021 he has been a guest editor for Frontiers in Computer Vision with a special issue on “Machine Vision for Assistive Technologies”. Since 2021, he serves as Associate Editor for The Visual Computer Journal. Since 2018, he has been involved in the collection, release, and maintenance of the EPIC-KITCHENS dataset series, and specifically in the egocentric action anticipation and in the action detection challenges. His research interests concern Computer Vision, Pattern Recognition, and Machine Learning, with focus on First Person Vision.
First Person (Egocentric) Vision for User-Centric AI Applications in Industrial Contexts
Wearable devices equipped with a camera, computing abilities, and some audio/visual feedback mechanism, such as Microsoft HoloLens and Google Glass, can perceive the world from the user’s point of view. This perspective, often referred to as “first person” or “egocentric”, offers exciting opportunities to both understand human perception and create AI algorithms capable of leveraging large amounts of human visual experience, which can be easily collected from this point of view. Indeed, such wearable devices offer an ideal platform to develop user-centric and personal AI applications with the aim to assist the users and augment their abilities. While the exploitation of Computer Vision in the industrial domain has historically relied on the analysis of images and video coming from fixed cameras watching the events from a “third person” point of view, first person vision is slowly making its way as a tool to improve the worker’s safety, augment their productivity, and make their training more efficient.
In this lecture, I will discuss the fundamental challenges and opportunities offered by egocentric vision, with a special focus on its application in industrial scenarios. I will cover the historical background and seminal works, present the main technological tools (including devices and algorithms) which can be used to develop AI applications and discuss challenges as well as open problems.