Vision-based AI in Social Signal Processing
Speaker: Dr. Marwa Mahmoud Abstract As technology became more widespread and personalised, fields like ‘affective computing’ and ‘social signal processing’ have emerged to investigate the possibility of building machines that can sense non-verbal signals from face and gesture and respond to natural human behaviour. Applications to these fields span different domains from: automotive industry to healthcare and animal welfare. In this talk, Marwa will talk about her work in building inference models that tackle challenging novel real-world problems, which are usually characterised by data scarcity and noisy signals from multiple modalities. Her work exploits multimodal feature representation, synthetic data generation, multimodal feature fusion, feature encoding, spatio-temporal feature representation, transfer learning, deep learning as well as extracting novel mid-level features for human behaviour understanding. Bio Dr. Marwa Mahmoud is a Junior Research Fellow of King’s College, Cambridge and an affiliated lecturer at the Department of Computer Science and Technology, University of Cambridge. She received her BSc and MSc degrees in Computer Science from the American University in Cairo. She then completed her PhD in 2015 in the Department of Computer Science and Technology, University of Cambridge. Her research interests lie in the field of computer vision and machine learning for affective computing, human behaviour understanding and social signal processing. She applied her research in the areas of automotive applications, healthcare, and animal welfare by working with collaborators across diﬀerent disciplines from academia and industry. In her spare time, she enjoys walking in the British countryside analysing faces of happy sheep.