Assistant professor, Kyushu University
Consumer grade RGB-D cameras such as the Kinect camera or the Asus Xtion pro camera have become the commodity tool to build dense 3D models of indoor scenes. The volumetric Truncated Signed Distance Function (TSDF) and its extensions have become popular and largely used for the task of 3D modeling using RGB-D cameras. However, this representation suffers several limitations due to its important memory footprint and unpractical manipulations. We propose a novel compact, flexible and accurate 3D representation based on parametric surfaces augmented by 2D deviation images. We propose a method to build in real-time animated 3D head models using a consumer-grade RGB-D camera. The facial motion is captured using a blendshape animation model while the geometric details are captured with a deviation image mapped over the template mesh. We demonstrate robust and high-fidelity simultaneous facial motion capture and 3D head modeling results on a wide range of subjects with various head poses and facial expressions.
Diego Thomas received his master's degree in Informatics and Applied Mathematics from the Ecole Nationale Superieure d'Informatique et de Mathematiques Appliquees de Grenoble (ENSIMAG), France in 2008. He received his Ph.D. from the Graduate University for Advanced Studies (SOKENDAI), Japan in 2012. He was a post-doc researcher at the National Institute of Informatics (Tokyo, Japan) from April 2012 to March 2015. From April 2015 to March 2017, he was a JSPS post-doc researcher at the Kyushu University (Fukuoka, Japan). His research interests are 3D images registration, 3D reconstruction and photometric analysis. From April 2017, he is assistant professor of Graduate School of Information Science and Electrical Engineering (ISEE), Kyushu University.
Professor, Kyushu University
Rin-Ichiro Taniguchi received the PhD degree in Engineering from Kyushu University, Japan, in 1986. In 1989, he became an associate professor of Interdisciplinary Graduate School of Engineering Sciences, Kyushu University. Since 1996, he has been a professor of Graduate School of Information Science and Electrical Engineering (ISEE), Kyushu University. From 2011 to 2014, he was the dean of ISEE, and, currently, he is the director of Research Institute for Information Technology, Kyushu University. His research interest includes computer vision, multimedia, cyber physical system, etc.
Professor, Paris-Est Marne-la-Vallée University.
Geometric algebras can be understood as a set of very intuitive tools to represent, construct and manipulate geometric objects. Among the various geometric algebras, this presentation will focus on those that may find some applications in computer sciences, namely Grassmann-Cayley algebra and the Conformal geometric algebra. This talk is an introduction in the domain, and thus no mathematical background is required. Moreover, this presentation aims to keep all concepts as simple and intuitive as possible. The plan of the talk will follow this guideline: first introduce the main concepts of geometric algebra such as outer product, geometric product or multivectors. We will see how Grassmann-Cayley algebra can be seen as an extension of the projective geometry massively used in computer graphics and computer vision. Then we will introduce some more powerful tools provided by conformal geometric algebra. In the second part of the talk, we will present our results on algebra implementations, their complexity in time and memory.
Vincent Nozick get his PhD in 2006 at Université Paris-Est Marne-la-Vallée (France) at the Laboratoire d'Informatique Gaspard Monge (LIGM). In 2006, he has been laureate of a Lavoisier fellowship for a post-doc position at Prof. Hideo Saito laboratory, Keio University. Then, he was hired as a tenured "maitre de conférences" at Université Paris-Est Marne-la-Vallée (France) from 2008. He served as the head master of the Imac engineer school from 2011 to 2013. He is currently in sabbatical at Japanese French Laboratory for Informatics (JFLI) at the National Institute of Informatics, Keio University and the University of Tokyo. In addition to computer vision applications, his current researches focus on both digital image forensics and geometric algebra.
Ph.D student, SOKENDAI (Graduate University for Advanced Studies)
We deal with the problem of fitting a discrete polynomial curve to 2D data in the presence of outliers. A discrete polynomial curve is defined as the set of integer points satisfying several inequalities. Finding a maximal inlier set from given data (integer points) that describes a discrete polynomial curve is equivalent with finding the feasible region corresponding to the set in the parameter space. When iteratively adding a data point to the current inlier set, how to update its feasible region is a crucial issue. We focus on how to track vertices of feasible regions in accordance with newly coming inliers. When a new data point is added to the current inlier set, a new vertex is obtained as the intersection point of an edge (or a face) of the feasible region for the current inlier set and a facet (or two facets) of the feasible region for the data point being added. Evaluating all possible combinations of an edge (or a face) and a facet (or two facets) is, however, computationally expensive. We propose an efficient computation in this incremental evaluation that eliminates combinations producing no vertices of the updated feasible region. This computation facilitates collecting the vertices of the updated feasible region. Experimental results demonstrate our proposed computation efficiently reduces practical running time.
Fumiki Sekiya recieved his master's degree in Engineering from Chiba University, Japan in 2014. From April 2014, he is a PhD student at SOKENDAI (Graduate University for Advanced Studies). His research interest includes discrete geometry.
Professor, Department of Information and Computer Science, Keio University
In this talk, I will present a future trend of visual processing in which the amount of images captured by various devices is extremely increasing. While machine learning has made drastic improvement based on the huge amount of image data, we need to achieve the innovative visualization technology using huge amount of image data. I believe that one of such trends is augmenting visual reality based on multi-dimensional visual information analysis for providing intuitive understanding using collection of various images. I will show some preliminary example research directions based on the collection images.
Hideo Saito received his Ph.D. degree in electrical engineering from Keio University, Japan, in 1992. Since then, he has been on the Faculty of Science and Technology, Keio University. From 1997 to 1999, he joined the Virtualized Reality Project in the Robotics Institute, Carnegie Mellon University as a visiting researcher. Since 2006, he has been a full professor in the Department of Information and Computer Science, Keio University. His recent activities for academic conferences include being Program Chair of ACCV2014, a General Chair of ISMAR2015, and a Program Chair of ISMAR2016. His research interests include computer vision and pattern recognition, and their applications to augmented reality, virtual reality, and human robotics interaction.
Post-doc researcher, National Institute of Informatics
In the age of permanent media exposure, video became our favorite mean of communication. Videos reflect our societies and can be seen from the lens of complex systems, gathering many types of entities, attributes and interactions over time, to which networks make a powerful mean of analysis. We can leverage on the visual potential of these structures to address multimedia analytics - the visual analytics of multimedia data. In this presentation, I will introduce different approaches we took to analyse and explore the NHK News 7 archive, namely with facetracking, multiplex networks, and graph signal processing.
After two years of studies at KAIST (Daejeon - South Korea), Benjamin joined his PhD program as a research engineer at the French National Audiovisual Institute (INA, Paris, France) and graduated from the University of Bordeaux (Bordeaux, France) in 2014. His work is at the intersection of two main disciplines. The first one is analysis and visualization of networks: he attempts explain complex phenomena at the heart of many disciplines such as biology, finances, or social sciences. The second one is multimedia analysis, which combines computer vision and data science to extract high level semantics from multimedia information. He joined the National Institute of Informatics in Japan as a JSPS fellow in 2014, where he now works on the analysis of the media landscape in Japanese television.
Senior researcher, Internet Initiative Japan
Understanding network health is essential to improve Internet reliability. For instance, detecting disruptions in peer and provider networks facilitates the identification of connectivity problems. Currently this task is time consuming for network operators. It involves a fair amount of manual observation because operators have little visibility into other networks. We leverage the RIPE Atlas measurement platform to monitor and analyze network conditions. We propose a set of complementary methods to detect network disruptions from traceroute measurements. A novel method of detecting changes in delays is used to identify congested links, and a packet forwarding model is employed to predict traffic paths and to identify faulty routers in case of packet loss. We present study cases demonstrating that the proposed methods detect real disruptions and provide valuable insights on the location and impact of identified events.
Romain Fontugne received his Master’s degree in Computer Science from Joseph Fourier University, France, in 2008. He received his Ph.D. degree from the Department of Informatics, the Graduate University for Advanced Studies (Sokendai), Japan, where he was working under the supervision of Kensuke Fukuda in Fukuda laboratory. From Nov. 2011 to Sep. 2013, Romain was a postdoctoral researcher (JSPS Fellowship) at the University of Tokyo, Esaki Laboratory. He worked from Oct. 2013 to Sep. 2015 at the National Institute of Informatics for the NECOMA project (Nippon-European Cyberdefense-Oriented Multilayer threat Analysis). His research interest are Internet traffic analysis, Internet security and traffic visualization. He is also a member of the WIDE project working on the MAWI archive.
Project assistant professor, the University of Tokyo
Cooperative Intelligent Transportation Systems (ITS) are systems where the vehicles, the roadside infrastructure, central control centres and other elements exchange information to achieve better road safety, traffic efficiency and comfort of the road users. These cooperative elements share the information by organising Vehicular Ad-hoc Network (VANET) or using the mobile network (e.g. 3G and LTE) and realises various cooperative ITS services. To promote wider deployment of such services, many stakeholders made consensus that the communication platform must be based on a common architecture and drive rapid international standardisation (e.g. ISO, ETSI, IEEE). Shortly, autonomous vehicles also benefit from such communication platforms by enforcing the perception, planing, and decision-making. However the most basic Vehicle-to-Vehicle (V2V) messages standardised in Europe, US and Japan suffer from same issues such as 1) unable to receive messages from the object without V2V transmitter, 2) message loss because of obstacle and wireless range, 3) vulnerable for malfunctioning and malicious node. For the solution, we propose an roadside-assisted V2V messaging system to support cooperative autonomous driving. We design the system based on the ITS Station architecture standardised in ISO/ETSI, working with any vehicle sensing technology. Moreover, we implement the prototype roadside system with a stereo vision for the vehicle sense. The prototype system is evaluated in a field test in the campus of the University of Tokyo. The results show that the proposed system significantly improves the coverage of V2V messaging while the system overhead is limited. The proposal is being integrated to our personal mobility autonomous vehicle system based on an open source software. The proposed system is designed to be technically compatible with 5G mobile edge computing.
Dr. Manabu Tsukada received his B.S. degree in Environmental and Information Studies from Keio University, Japan, in 2005 and his M.S. degree in Media and Governance from Keio University, Japan, in 2007. He worked in IMARA Team, INRIA, France during his Ph.D course and obtained his Ph.D. degree from Centre de Robotique, Mines ParisTech, France, in 2011. During his pre and postdoc research stages he has participated in multitude of international projects in the networked ITS area, such as GeoNet, ITSSv6, SCORE@F, CVIS, Nautilus6 or ANEMONE. He has been serving as board member of the WIDE Project since 2014. Regarding scientific dissemination, Dr. Tsukada has published a number of conference and journal papers in the ITS and computer networks areas, and his appearance in IPv6 related events and initiatives is common. He is currently a project assistant professor at Graduate School of Information Science and Technology, the University of Tokyo, Japan. His research interests are mobility support for next generation Internet (IPv6) and communications for intelligent vehicles.
JSPS Post-doc researcher, the University of Tokyo
The Internet eXchange Points (IXP) are essential for the Internet evolution as they empower high bandwidth low latency and inexpensive local traffic peering as opposed to transit traffic. On the other side, Software Defined Networking, or SDN for short (and its principal realization, the OpenFlow protocol) enables more network programmability to control network behavior via open interfaces, as opposed to the legacy closed-box solutions and proprietary-defined interfaces. This talk is about designing, deploying and operating SDN IXP with the experience of TouSIX and NSPIXP-3 at Osaka.
Marc Bruyere started his career in 1996 working for Club-Internet.fr, and for Cisco, Vivendi Universal, Credit Suisse First Boston, Airbus/Dimension Data, Force10 Networks, and Dell. He received is Ph.D degree from the LAAS CNRS, his thesis is about Open Source OpenFlow SDN for IXPs. He designed and deployed the first European OpenFlow IXP fabric for the TouIX. Now he is a PostDoc at the University of Tokyo.
Assistant professor, Kyushu University
Vehicle counting is one of the fundamental tasks in the ITS (intelligent transportation systems). To retrieve realtime traffic data, automatic vehicle counters have been deployed. However, the deployment of the automatic vehicle counters is limited to high traffic roads because of their high installation and maintenance costs. We are developing a low-cost acoustic vehicle sensor that utilizes a stereo microphone at a roadside. In this presentation, I present design basics as well as challenges of the acoustic vehicle sensor.
Shigemi Ishida received the B.E. degree in electrical engineering from the Shibaura Institute of Technology, Tokyo, Japan, in 2006, and the M.S. degree and the Ph.D. degree in electrical engineering and information system from the University of Tokyo, Japan, in 2008 and 2012, respectively. He was a Visiting Scholar with the Department of Computer Science and Engineering, University of Minnesota, Minneapolis, MN, USA, in 2013. He is currently an Assistant Professor with the Faculty of Information Science and Electrical Engineering, Kyushu University, Fukuoka, Japan. He is currently focused on localization systems and sensing technologies for better understanding of the world around us. His research interests include wireless sensor networks, low-power wireless communications, localization systems, cross-technology communications, and intelligent transportation systems.