Multiview video coding using cuboid data compression

Presenters:

Prof. Manzur Murshed (Monash University) and Dr. Manoranjan Paul (Charles Sturt University)

Abstract

Capturing a scene using multiple cameras is expected to provide the necessary interactivity in the three-dimensional (3D) space to satisfy end-users' demands for observing objects and actions from different angles and depths. Proliferation of video games has inevitably created a natural demand to introduce similar interactivity while watching a TV broadcast. Traditional single-view video coding architecture can only offer passive viewing experience and hence, a new multi-view video coding (MVC) paradigm is introduced to enable efficient encoding of sequences captured simultaneously from multiple cameras using a single video stream. MVC covers a wide range of active viewing experience, including stereoscopic (two-view) video (popularly known as 3DTV), free viewpoint television (2D view from any viewing angle can be generated from 3D scene modelling) and multi-view 3DTV. Considering the significant overlapping of the views and, more importantly, the availability of a rich set of relations on the geometric properties of a pair of views from camera properties, known as the epipolar geometry, joint encoding/decoding of views can achieve significant compression that independent simulcasting of views. Taking advantage of the new inter-view correlation channel, in addition to traditionally exploited intra-frame and inter-frame correlation channels, is quite intricate. It makes some of the well-established video coding practices and standards more challenging and introduces new interactivity quality metrics such as random access view delay. Nevertheless, MVC also opens up opportunities for addressing some of the most stubborn video coding problems. For example, multiple views with overlapped area of coverage can offer a natural solution to the occlusion/de-occlusion problem, which often leads to incorrect object recognition and tracking in video surveillance applications.

This tutorial will introduce multiple view geometry, present a systematic review of advancements in MVC, and outline our recent high-interactivity MVC framework using cuboid (volumetric) data compression and dynamic panorama background modelling.

Presenter Bio(s)

Prof. Manzur Murshed (Monash University, Australia)

Manzur Murshed received the B.Sc.Eng. (Hons.) degree in computer science and engineering from Bangladesh University of Engineering and Technology (BUET), Dhaka, Bangladesh, in 1994 and the Ph.D. degree in computer science from the Australian National University, Canberra, Australia, in 1999. He is currently an Associate Professor and Head of Gippsland School of Information Technology, Monash University, Australia and was one of the founding directors of the Centre for Multimedia Computing, Communications, and Applications Research (MCCAR). His major research interests are in the fields of video technology, information theory, wireless communications, distributed computing, and security & privacy. He has so far published 170 refereed research papers and received more than $1 million nationally competitive research funding, including three Australian Research Council Discovery Project grants in 2006, 2010, and 2013 on video coding & communications and a large industry grant in 2011 on secured video conferencing. To date he has successfully supervised 18 PhD students. He is an Editor of International Journal of Digital Multimedia Broadcasting and has had served as an Associate Editor of IEEE Transactions on Circuits and Systems for Video Technology in 2012 and as a Guest Editor of special issues of Journal of Multimedia in 2009-2012. He received a University Gold Medal from BUET in 1994, the inaugural Early Career Research Excellence award from the Faculty of Information Technology, Monash University in 2006, and the Vice-Chancellor's Knowledge Transfer Award (commendation) from the University of Melbourne in 2007. He is a senior member of IEEE.

Dr. Manoranjan Paul (Charles Sturt University, Australia)

Manoranjan Paul received B.Sc.Eng. (hons.) degree in Computer Science and Engineering from Bangladesh University of Engineering and Technology (BUET), Bangladesh, in 1997 and PhD degree from Monash University, Australia in 2005. He was a Lecturer and Assistant Professor in the department of Computer Science and Engineering, Ahsanullah University of Science and Technology, Bangladesh from 1997 to 2001. He was a Post-Doctoral Research Fellow in the University of New South Wales, ADFA, in 2005~2006, Monash University, in 2006~2009, and Nanyang Technological University, in 2009~2011. He has joined in the School of Computing and Mathematics, Charles Sturt University (CSU) at 2011. Currently he is a Senior Lecturer and Associate Director of the Centre for Research in Complex Systems in CSU. His major research interests are in the fields of video coding, computer vision, and EEG Signals analysis. He has published more than 65 refereed publications. Dr. Paul regularly published journal articles in the IEEE Transactions on Image Processing, IEEE Transactions on Circuits and Systems for Video Technology, and IEEE Transactions on Multimedia, which are considered as the top ranked journals in the image processing, video technology, and multimedia fields respectively. He organized a special session on "New video coding technologies" in IEEE ISCAS 2010. He is a keynote speaker on "Vision friendly video coding" in IEEE ICCIT 2010.

Dr Paul is a full member of the IEEE ('03) and senior member of ACS ('05). Dr. Paul is an Associate Editor of EURASIP Journal on Advances in Signal Processing (JASP). Dr. Paul has served as a guest editor of Journal of Multimedia from 2008-2013. Currently Dr. Paul is an Editor of the International Journal of Engineering and Industries (IJEI) and the International Journal of Soft Computing, Mathematics and Control (IJSCMC). He has been on Technical Program Committees of more than 20 international conferences. Dr. Manoranjan Paul obtained Research Excellence Award 2011 from School of Computing and Mathematics, CSU. Dr. Paul obtained Australian Research Council (ARC) Discovery Project Grant with Professor Manzur Murshed. So far Dr. Paul obtained more than $1,000,000 competitive grant money.

Back to top

Satellite Image Analyses for Marine and Coastal Applications

Presenter:

Dr. Vincent Lyne (CSIRO Marine and Atmospheric Research, Australia)

Abstract

Satellite images are increasingly relied upon for management, monitoring and mapping a variety of marine and coastal water properties, and to investigate changes over time. The first part of this tutorial will cover an application of satellite data to understand the arrangement of water mass types and their structural properties in the ocean. We'll go through some basic spatial analytical techniques and image feature analyses for noisy data, and we'll draw some lines and boundaries on water. In the second part, we'll examine a time sequence of images to explore various facets of the regime shift occurring in the Tasman Sea. In the last part of the tutorial we'll debate strategies for analysing complex dynamically changing patterns. What are the strategic approaches for gaining deep insight; in essence a "pattern recognition strategy" challenge to pattern recognition? Is it better always to have greater dynamical resolution of the problem? Or, is it better to use simpler approaches - why and what are these? Are there generic approaches and lessons we can apply to any pattern recognition problem? Come prepared to challenge existing paradigms, and to be challenged.

Presenter Bio

Profile

Dr. Vincent Lyne (CSIRO Marine and Atmospheric Research, Australia)

Research Scientist with CSIRO Marine and Atmospheric Research with experience in hydrology, rainfall-runoff models, sediment transport models, groundwater resource assessments, and named on the "Lyne Hollick algorithm" used globally for water management. A key contributor to activities that underpin Australia's marine regional planning program, and contributed to the development of whole-of-system approaches and appropriate analytical tools to facilitate participatory planning projects on sustainable livelihoods in developing countries. Experienced in systems-based approaches for making intelligent decisions on complex environmental and socio-economic problems. An Australian Research Council Reviewer and part of the Nobel Peace Prize to the International Panel on Climate Change (IPCC). Has been an official Australian delegate to the Chinese Academy of Sciences and the London Dumping Convention.

Key Research Skills

• Hydrology and rainfall-runoff analyses
• Time-series analyses, forecasting and regime shift identification
• GIS mapping, spatial analyses, remote sensing, spatial characterisation
• Oceanography, sediment transport, circulation, water quality, flushing
• Fisheries, marine and coastal biodiversity
• Ecosystem characterisation and assets-threats analyses
• Socio-economic typologies and dependence on ecosystem services
• Risk (oil spills, climate change, human use) identification, vulnerabilities, trends, triggers/thresholds, adaptations
• Cross scale dependencies and network interaction analyses

Back to top