Abstract:
Among the various flavors of shape analysis, shape approximation methods offer versatile frameworks to capture the essence of a 3D shape with a compact set of numbers, sometime called proxy. In this talk, I will give an overview of recent results we achieved regarding shape approximation of 3D surfaces, including Sphere Meshes, which help simplifying complex objects at extreme levels, providing multi-resolution spatial interfaces for interactive modeling systems and automatic processing chains. I will also discuss constrained approximation with Bounding Proxies, which leverage mathematical morphology for generating simplified meshes that contain their input and can be used for physical simulation, cage-based deformation or level-of-detail rendering. Finally I will extend the discussion to the general graphics pipeline and give insights on how shape proxies can be key components for numerous applications in the era of massive 3D models and smart computer graphics tools.

About the Speaker:
Prof. Dr. Tamy Boubekeur is a Full Professor in Computer Science at Telecom ParisTech, Paris-Saclay University, Paris, France, where he leads the 3D computer graphics group. His research focuses on shape modeling, image synthesis and interactive systems. From 2004 to 2007, he was a member of INRIA Bordeaux, France, and a regular invited scientist at the University of British Columbia, Vancouver, Canada. He received his PhD in Computer Science from the University of Bordeaux in 2007. Then, he joined TU Berlin in Germany as an Associate Researcher. In 2008, he joined the "Image Data Signal" Department of at Telecom ParisTech as an Associate Professor where he founded the Computer Graphics Group. He received his Habilitation à Diriger des Recherches (HDR) in Computer Science from University ParisSud in 2012. He became Full Professor at Telecom ParisTech in 2013. Tamy has published a number of articles in top-ranked international conferences and journals in computer graphics and computer vision, and received several international scientific awards. He has been member of a number of international scientific committee, including SIGGRAPH, SIGGRAPH Asia, EUROGRAPHICS, EGSR, SGP, HPG, SMI and I3D. He is also Allegorithmic’s Chief Scientist and holds a second, part-time Professor position at Ecole Polytechnique, Paris, France. More information at www.telecom-paristech.fr/~boubek.

Abstract:
Single-cell analysis through mass cytometry has become an increasingly important tool for immunologists to study the immune system in health and disease. Mass cytometry creates a high-dimensional description vector for single cells by time-of-flight measurement. In this talk we will discuss several hierarchical approaches to the interactive exploration of large single cell data using a combination clustering and t-Distributed Stochastic Neighborhood Embedding (t-SNE) as well as the recently introduced Hierarchical Stochastic Neighborhood Embedding (HSNE). Based on the application to a study on gastrointestinal disorders we show hat HSNE efficiently replicates previous observations and identifies rare cell populations that were previously missed. Finally we will discuss CyteGuide, a tool to guide the exploration of HSNE hierarchies.

About the Speaker:
Thomas Höllt received the Diplom (MSc) in computational visualistics from the University of Koblenz-Landau, Germany, in 2008, and the PhD in computer science from the King Abdullah University of Science and Technology, Saudi Arabia, in 2013. He holds positions as an Assistant Professor in the Computational Biology Center (https://www.lcbc.nl) at the Leiden University Medical Center and as a research fellow at Delft University of Technology.

Abstract:
For mixed reality applications, where reality and virtual reality are spatially merged and aligned in interactive real-time, we propose a voxel representation as a rendering and interaction method for the near future. We show that voxels—gap-less volumetric pixels in a regular grid in space—allow for an actual user experience of a mixed reality environment. We demonstrate that a low fidelity voxel representation can lead to sufficient levels of presence and co-presence.
We argue the case for voxels by (1) conceptually defining and illustrating voxel-based mixed reality, (2) presenting a low resolution and fully functioning prototype, (3) empirically exploring the user experience, (4) describing the computational feasibility, and (5) finally discussing future directions for voxel-based mixed reality. If everything is based on voxels, even if coarse, visual coherence is achieved inherently.

About the Speaker:
Holger Regenbrecht is a Professor at the Department of Information Science at the University of Otago. He obtained his doctorate from Bauhaus University and has been working in the fields of virtual and mixed Reality for 20 years. He was initiator and manager of the Virtual Reality Laboratory at Bauhaus University Weimar (Germany) and the Mixed Reality Laboratory at DaimlerChrysler Research and Technology (Ulm, Germany). Now he co-leads the Human-Computer Interaction Lab at the University of Otago.
Dr Regenbrecht's research interests include human-computer interaction (HCI), applied computer science and information technology, (collaborative) augmented reality, 3D teleconferencing, psychological aspects of mixed reality, three-dimensional user interfaces (3DUI) and computer-aided therapy and rehabilitation. In those areas he has published one hundred peer-reviewed articles. His current work focuses on translational ICT research, in particular for health and wellbeing and on understanding computer-mediated realities.
He is a member of several international professional groups and serves as an editorial board member, reviewer and auditor for a number of conferences, journals and institutions. Holger is the current head of the Information Science department.