Seminar by Charalambos Poullis (Cyprus University of Technology)
Speaker: Charalambos Poullis
Cyprus University of Technology
Title: Virtual Environments: Creation & Applications
Date: Monday, February 16th, 2015
Time: 10:30AM - 12PM
In recent years there has been an increasing demand for applications which employ miniature representations of the real world to recreate realistic and immersive virtual environments. Many applications ranging from computer vision/graphics and virtual reality, to Geographical Information Systems have already successfully used real world representations derived from the combination of multi-sensory data captured from aerial or satellite imagery and LiDAR (Light Detection and Ranging) scanners. However, despite their widespread and successful application, the creation of such realistic 3D content remains a complex, time-consuming, expensive and labor-intensive task. In fact, the creation of models is still widely viewed as a specialized art, requiring personnel with extensive training and experience to produce useful models.
In this talk, I will present novel algorithms developed to address the historically-difficult problems in creating large-scale (city size) virtual environments from active sensor data, including rapid extraction and modeling of geometric models, re-production of high-quality scene textures, and fusion and completion of the geometry and texture data in order to produce photorealistic 3D scene models. More specifically, I will introduce a complete framework for the automatic modeling from pointcloud data which does not make any particular assumptions about the data e.g. spatial coherence, symmetry constraints, etc, and whose performance and quality of the results is independent of the size of the data. As part of this framework, I will also present the robust unsupervised clustering algorithm P2C and a fast boundary refinement process based on global optimization. Next, I will present a novel vision-based system for the automatic detection and extraction of complex road networks from various sensor resources. Uniquely, the system is an integrated solution that merges the power of perceptual grouping theory and optimized segmentation techniques into a unified framework to address the challenging problems of geospatial feature detection and classification. Finally, I will present the results of a series of Virtual Reality case studies, where users are immersed in, and interact with, virtual environments using a CAVE system. I will conclude the presentation by proposing a research agenda for the future.
Charalambos Poullis received his B.Sc. in Computing Information Systems with First-class Honors from the University of Manchester, UK in 2001, his M.Sc. in Computer Science with specialization in Multimedia and Creative Technologies from the University of Southern California (USC) in 2003 and his PhD in Computer Science from USC in 2008.
Charalambos' research interests lie at the intersection of computer vision and computer graphics. More specifically, he is interested in conducting fundamental and applied research in the following areas: data acquisition & 3D reconstruction, photorealistic rendering, feature extraction & classification, virtual & augmented reality. He is a member of the ACM and IEEE and has served as a regular reviewer in several conferences and journals.
He currently holds a tenure-track Assistant Professor position with the Department of Multimedia and Graphic Arts at the Cyprus University of Technology where he also serves as the Director of the Immersive and Creative Technologies (ICT) lab.