Rate-Distortion Optimization in Video/Image Compression: From Temporal Dependency Formulation to Learning-based Modeling

Abstract

The past decades have witnessed remarkable advancements in image/video coding techniques and their diverse applications, where rate-distortion optimization (RDO) plays a crucial role in maximizing coding efficiency. In the traditional block-based hybrid coding framework, the extensive utilization of spatial-temporal prediction makes the processing of each coding unit highly dependent and entangled, rendering the global RDO problem particularly challenging. One feasible solution is to delicately formulate the spatial-temporal dependencies accurately. I will first discuss achieving temporally dependent RDO in one pass coding, and present our RDO work on top of various video coding standards, from H.264/AVC to H.265/HEVC, and to the latest Versatile Video Coding (H.266/VVC). On the other hand, leveraging the powerful nonlinear transformation and entropy modeling capabilities of neural networks, end-to-end learning-based approaches have emerged as a highly promising alternative for achieving high efficiency image/video compression. In the second part, I will touch some work on end-to-end learning-based image compression, particularly highlighting our recent advancements in optimizing the entropy model.

Biography

Ce Zhu (IEEE/Optica Fellow) is currently a Professor at the University of Electronic Science and Technology of China, Chengdu, China. His research interests include image/video coding and communications, 3D video, visual analysis, perception and applications. He has served on the editorial boards of a few journals, including as an Associate Editor of IEEE Transactions on Image Processing (2016-2020). He has also served as a Guest Editor of a few special issues in international journals, including as a Guest Editor in the IEEE Journal of Selected Topics in Signal Processing (2020). He was an APSIPA Distinguished Lecturer (2021-2022), and also an IEEE Distinguished Lecturer of Circuits and Systems Society (2019-2020). He is serving as the Chair of IEEE ICME Steering Committee (2024-2025), and the Chair of IEEE Chengdu Section. He is a co-recipient of multiple paper awards at international conferences, including the Best Demo Award in IEEE MMSP 2022, and the Best Paper Runner Up Award in IEEE ICME 2020. For more information, please visit his homepage at http://www.avc2-lab.net/~eczhu

AI and Cognitive Health

Abstract

With the global population ageing rapidly, a key health concern lies in Neurocognitive Disorders (NCD), also known as dementia -- a common form being Alzheimer's Disease (AD). NCDs are particularly prominent in older adults, which has an insidious onset followed by gradual, irreversible deterioration in cognitive domains (memory, communication, etc.). Thus the screening NCD is crucial for timely intervention to slow down disease progression. We will present our work in the use of spoken language for assessing cognitive health, including a carefully designed speech data collection protocol that has contextual relevance to the linguistic environment in Hong Kong, Macao and the Greater Bay Area, the development of AI-enabled speech and language analytics approaches, and the use of different acoustic and linguistic features for the AD detection. Experimental results demonstrate the feasibility of automating cognitive health monitoring with the use of speech and language technologies, which offers the advantages of accessibility, non-invasiveness and affordability. This work is funded by the HKSAR Government's Theme-based Research Scheme.

Biography

Helen Meng is Patrick Huen Wing Ming Professor of Systems Engineering & Engineering Management at The Chinese University of Hong Kong. She received all her degrees from MIT and joined CUHK in 1998. She is the Founding Director of the Microsoft-CUHK Joint Laboratory for Human-Centric Computing and Interface Technologies in 2005, which has been recognized as a Ministry of Education of China (MoE) Key Laboratory since 2008. In 2006, she founded the Tsinghua-CUHK Joint Research Centre for Media Sciences, Technologies and Systems and has served as its Director. In 2013, she helped establish the CUHK Stanley Ho Big Data Decision Analytics Research Center and serves as its Founding Director. She served as former Associate Dean (Research) of Engineering (2006-2010), and former Chairman of the Department (2012-2018).
Helen’s professional services include former Editor-in-Chief of the IEEE Transactions on Audio, Speech and Language Processing, and a member of the IEEE Board of Governors. She has served or is serving as a member of the Advisory Panel of the Hong Kong Science and Technology Park Corporation, the review panels of the Swedish Research Council European Research Infrastructure Initiative, and the National Centres of Competence in Research of the Swiss National Science Foundation. She is a member of the HKSAR Government’s Steering Committee on eHealth Record Sharing, Convenor of the Engineering Panel HKSAR Government’s Competitive Research Funding Schemes for the Self-financing Degree Sector, member of the Hong Kong/Guangdong ICT Expert Committee and Coordinator of the Working Group on Big Data Research and Applications, Council membership of the Open University of Hong Kong, member of the Research Grants Council, former Council Member of the Hong Kong Productivity Council, former member of the HKSAR Government’s Digital 21 Strategy Advisory Committee, and Chairlady of the Working Party of the Manpower Survey of the Information Technology Sector (undertaken by the Hong Kong Census and Statistics Department) for 2014-2017.
Helen is a recognized scholar in her field. She leads the interdisciplinary research team that received the first Theme-based Research Scheme Project in Artificial Intelligence in 2019. Her recent awards include 2019 IEEE Signal Processing Society Leo L Beranek Meritorious Service Award, 2018 CogInfoComm Best Paper Award, 2017 Outstanding Women Professional Award (one of 20 since 1999), 2016 Microsoft Research Outstanding Collaborator Award (one of 32 academics worldwide), 2016 IBM Faculty Award, 2016 IEEE ICME Best Paper Award, 2015 ISCA Distinguished Lecturer, 2015 HKCS inaugural Outstanding ICT Women Professional Award and 2012 Asia-Pacific Signal and Information Processing Association (APSIPA) inaugural Distinguished Lecturer. Prior to that, she has also received such awards as the CUHK Faculty of Engineering Exemplary Teaching Award, Young Researcher Award and Service Award; APSIPA Best Oral Paper Award, and 2009 Ministry of Education Higher Education Outstanding Scientific Research Output Award in Technological Advancements. She has delivered numerous invited and keynote talks, such as IEEE SIDAS 2016, ASTRI-HPE Conference 2016, Internet Economy Summit 2017, GMIC 2017, INTERSPEECH 2018 Plenary, SIGDIAL 2019 Keynote, etc. She is a Fellow of the Hong Kong Computer Society, Hong Kong Institution of Engineers, International Speech Communication Association and IEEE.