Events

Affect Analysis in-the-wild: Valence-Arousal, Expressions, Action Units and a Unified Framework, seminar by Dr Dimitrios Kollias

Dr Dimitrios Kollias, Lecturer in Computer Science, Queen Mary University of London

Date: 16 March 2022 14:30 - 15:00

Location: online, Teams: https://teams.microsoft.com/l/meetup-join/19%3a4dba808159354...

Affect Analysis in-the-wild: Valence-Arousal, Expressions, Action Units and a Unified Framework

Dr Dimitrios Kollias, Lecturer in Computer Science, Queen Mary University of London

Abstract. Affect recognition based on subjects' facial expressions has been a topic of major research in the attempt to generate machines that can understand the way subjects feel, act and react. In the past, due to the unavailability of large amounts of data captured in real-life situations, research has mainly focused on controlled environments. This has changed in the last few years and some large in-the-wild databases have been developed. Moreover, deep learning has emerged as a means to solve visual analysis and recognition problems. In this talk I will present significant contributions for affect analysis and recognition in-the-wild.
Affect analysis and recognition can be seen as a dual knowledge generation problem, involving: i) creation of new, large and rich in-the-wild databases and ii) design and training of novel deep neural architectures that are able to analyse affect over these databases and to successfully generalise their performance on other datasets. The talk focuses on the design of two classes of deep neural networks trained with these databases. The first class refers to uni-task affect recognition; the second class refers to multi-task learning of all main behavior tasks, i.e. valence-arousal estimation; basic expression classification; facial Action Unit detection.

Bio. Dimitrios Kollias, FHEA, PG-Cert holder and IEEE member, is currently a Lecturer in AI at EECS. He has been the recipient of the prestigious Teaching Fellowship of Imperial College London. He has obtained the Ph.D. from Imperial College London, where he was a member of the iBUG group. He has been a Senior Lecturer in Artificial Intelligence at the University of Greenwich. He has worked as Research Scientist at Deep Render Ltd on deep learning based image/video compression; he has been Consultant at FaceSoft Ltd on deep learning based affect recognition and generation; he has been Consultant at Cogitat Ltd on decoding people's music-listening brain states; he has collaborated with Realeyes Company on machine learning based affect and attention methodologies. Two patents on image analysis have been generated and filed. He has published his research in the top journals and conferences on machine learning, perception and computer vision, such as IJCV, CVPR, ICCV, ECCV, BMVC, IJCNN and ECAI, also serving as a reviewer in them. He is the Chair of the ABAW and MIA-COV19D Competitions and Workshops in IEEE CVPR 2022, ICCV 2021 and IEEE FG 2020. His research interests span the areas of machine and deep learning, domain adaptation, computer vision, affective computing and medical imaging.

Updated by: Ildar Farkhatdinov