Call for Participation

 
 

First ACM/IEEE International Conference on Distributed Smart Cameras (ICDSC-07)

September 25-28, 2007
Vienna, Austria

Tutorials:

Andrea Cavallaro, Queen Mary University of London, UK

I. Smart Cameras: Algorithms, Evaluation and Applications


Abstract
This tutorial is a complete and balanced review of the subject having a broad scope and coverage of algorithms and applications of smart cameras. The tutorial begins with an introduction and extensive review of various algorithms for single and multi-modal object detection and tracking, from basic principles to various steps required to obtain accurate object extraction using single as well as distributed smart cameras. This part of the tutorial includes data analysis methods for quantitative evaluation of smart camera results. Next, the presentation covers a range of smart camera applications, including event recognition, behaviour analysis, multi-camera calibration and scene reconstruction. Relevant output formats for smart cameras are also discussed. Computer implementations of dedicated software packages and their research applications are described and illustrated with some useful links and examples. Various quantitative assessments for comparison of smart camera results will be presented including well-known figures of merit and protocols used in international evaluation campaigns. The final part of the tutorial includes a discussion of relevant ethical and societal issues related to the development and use of smart cameras in public and private spaces as well as an in-depth discussion of open research issues. We will also provide a series of relevant links to available code for smart cameras.

Click here for an example video.

Lecturer Biography
Dr. Andrea Cavallaro is lecturer (UK equivalent to Assistant Professor in North America) at the Department of Electronic Engineering, Queen Mary, University of London (QMUL). He was workpackage leader for the EU projects ACTS Modest and IST art.live and is Principal Investigator in a number of UK Research Council and industry-sponsored projects. Dr. Cavallaro was a Research Fellow with British Telecommunications (BT) in 2004/2005; he was awarded the Drapers' Prize for the development of Learning and Teaching in 2004; an e-learning Fellowship in 2006; and the Royal Academy of Engineering teaching Prize in 2007. He is co-author of the papers on target tracking winner of the student paper contest at the IEEE ICASSP in 2005 and 2007. Dr. Cavallaro is an elected member of the IEEE Signal Processing Society, Multimedia Signal Processing Technical Committee, General Chair of the IEEE International Conference on Advanced Video and Signal based Surveillance (AVSS 2007), Chair of the 2007 BMVA symposium on Security and Surveillance; Technical co-chair of the European Signal Processing Conference (EUSIPCO 2008), Guest Editor of the Special Issue on 'Multi-sensor object detection and tracking', Signal, Image and Video Processing Journal (Springer) and co-Guest Editor of the Special Issue on 'Video Tracking in Complex Scenes for Surveillance Applications', Journal of Image and Video Processing. He is a member of the organizing/technical committee for several international conferences, and he is author of more than 60 papers, including 5 book chapters.

Click here for lecturer's website.

Bjoern Gottfried, University of Bremen, Germany

II. Ambient Intelligence and the Role of Spatial Reasoning: Smart Environments with Smart Cameras


Abstract
A short historial overview starts the tutorial, introducing what AmI is all about, showing different subfields of this area. Differences are discussed which are used for classifying applications, such as ubiquitous, pervasive, wearable, and mobile computing. In the context of AmI further terms are clarified, in particular to explain what smart, intelligence, ambience and some other catchwords might mean or could mean in future. An outline of the tutorial shows how the tutorial gives an overview of some important ingredients of AmI systems but simultaneously it is argued that it lies in the complexity of the topic that the tutorial has to restrict itself to some more or less specific subareas which can be discussed more deeper.

Contrasting the whole field from other technological areas which have become prominent over the last decades new challenges are pointed out. For example, another such field is the semantic web which is just to mention as it is situated in virtual space whereas AmI systems are to be realised in real space. This brings in new challenges since concepts like environment, location, motion, and real objects enter the scene. This is why the tutorial in particular discusses AmI systems from the point of view of spatial and also a little bit temporal reasoning, both areas fundamental for any AmI system to properly work.

Since the tutorial takes place on the International Conference on Smart Cameras particular emphasis is put on what role smart cameras might play in AmI systems or already play. It is therefore that example systems are presented which make use of camera setups in AmI environments, such as in the healthcare area. Furthermore, smart cameras form an important class of sophisticated sensors using which spatial information of objects acting in environments can be recorded. Means for using this information are therefore shown, for example, by demonstrating how smart cameras cooperate in an AmI system as to allow for more information when working together.

The tutorial closes by summarising the particular importance of the methods discussed. But it is also pointed out where current problems remain, and new challenges and opportunities lie. Besides a broad overview of the field, participants will especially and finally have an idea on methods, problems, and possibilities for why and how spatial concepts (environments, locations, motion, etc.) play an important role in the context of AmI systems.

Lecturer Biography
Björn Gottfried studied software engineering at the University of Applied Sciences in Hamburg. After some experience in the software industry he studied computer science at the University of Bremen. He received his doctoral degree by 2005 in the context of spatial reasoning. Currently he works at the Centre for Computing Technologies at the University of Bremen as a research scientist and lecturer in the context of AI, in particular about image processing, and spatial and diagrammatic reasoning. Over the last ten years he published over thirty papers, mainly about spatial and temporal reasoning, image processing, and ambient intelligence. He is pc member in several workshops about ambient intelligence and related fields and organises a workshop on behaviour monitoring and interpretation.

Click here for lecturer's website.

Richard Radke, Rensselaer Polytechnic Institute, USA

III. Multiview Geometry for Camera Networks


Abstract
This tutorial will introduce the basics of multiview geometry in computer vision, including: camera models, projective geometry, auto-calibration, epipolar geometry and the fundamental matrix, 3D reconstruction, N-view geometry, structure from motion, and bundle adjustment. We will also discuss important related topics, including feature detection, description, and matching. We will then present a more detailed discussion of how to obtain the accurate and globally consistent self-calibration of a distributed camera network, in which cameras and processing nodes may be spread over a wide geographical area, with no centralized processor and limited ability to communicate a large amount of information over long distances. First, we describe how to estimate the vision graph for the network, in which each camera is represented by a node, and an edge appears between two nodes if the two cameras jointly image a sufficiently large part of the environment. We propose an algorithm in which each camera independently composes a fixed-length message that is a lossy representation of a subset of detected features, and broadcasts this "feature digest" to the rest of the network. Each receiver camera decompresses the feature digest to recover approximate feature descriptors, robustly estimates the epipolar geometry to reject outliers and grow additional matches, and decides whether sufficient evidence exists to form a vision graph edge. Second, we present a distributed camera calibration algorithm based on belief propagation, in which each camera node communicates only with its neighbors in the vision graph. The natural geometry of the system and the formulation of the estimation problem give rise to statistical dependencies that can be efficiently leveraged in a probabilistic framework. The camera calibration problem poses several challenges to information fusion, including missing data, overdetermined parameterizations, and non-aligned coordinate systems. We demonstrate the accurate and consistent performance of the vision graph generation and camera calibration algorithms using a simulated 60-node outdoor camera network.

Lecturer Biography
Richard J. Radke received the B.A. degree in mathematics and the B.A. and M.A. degrees in computational and applied mathematics, all from Rice University, Houston, TX, in 1996, and the Ph.D. degree from the Electrical Engineering Department, Princeton University, Princeton, NJ, in 2001. For his Ph.D. research, he investigated several estimation problems in digital video, including the synthesis of photorealistic "virtual video", in collaboration with IBM's Tokyo Research Laboratory. He has also worked at the Mathworks, Inc., Natick, MA, developing numerical linear algebra and signal processing routines.

He joined the faculty of the Department of Electrical, Computer, and Systems Engineering, Rensselaer Polytechnic Institute, Troy, NY, in August, 2001, where he is now an Associate Professor. He is also associated with the National Science Foundation Engineering Research Center for Subsurface Sensing and Imaging Systems (CenSSIS). His current research interests include deformable registration and segmentation of three- and four-dimensional biomedical volumes, machine learning for radiotherapy applications, distributed computer vision problems on large camera networks, and modeling 3D environments with visual and range imagery. Dr. Radke received a National Science Foundation CAREER Award in 2003, and is a member of the 2007 DARPA Computer Science Study Group.

Click here for lecturer's website.

Wilfried Elmenreich, Vienna University of Technology, Austria

IV. Real-time Sensor Networks for Smart Cameras: Communication, Data Processing and Applications


Abstract
This tutorial introduces the topic of real-time smart sensor networks and its application to smart camera networks from different perspectives. The tutorial splits up in three parts, where the first part discusses the low-level issues like clock synchronization and communication, the second part treats the data processing and refinement operations, and the third part gives two application scenarios for networks of smart cameras.

In the first part of the tutorial, the basic concepts of networking smart sensors for real-time applications are discussed: The concept of a sparse time base enables a consistent view of events and supports the identification of a global system state. Clock synchronization is necessary to agree on a common global time within the network. Establishing a global time requires some communication and computation effort, but simplifies the interpretation of measurements and timestamps. A concise description of interfaces in both the value and the temporal domain supports a two-level component-based design approach supporting certification and reuse. For dependability, fault-tolerance can be implemented by redundancy and fault containment concepts.

In the second part, we discuss data processing types within the network by the example of a sensor fusion process. Thus, data from several sensors can be combined to improve the result in terms of accuracy, completeness, or creating emergent information. For example, the images from two cameras with overlapping views may be combined to render depth information or to reduce noise.

In the third part, we discuss the requirements for several application scenarios of sensor networks for image sensors. For two application scenarios, example network architectures are presented: a time-triggered network for networking cameras supporting driver assistance functions in a car, and a wireless network using ZigBee Technology and low power nodes for video surveillance.

Lecturer Biography
Wilfried Elmenreich is an assistant professor at the Institute of Computer Engineering at Vienna University of Technology in Austria. He studied at the Engineering School for Electrotechnics and Control in Weiz, Styria, Austria, and graduated at the Vienna University of Technology. He received a Master's degree in computer science in 1998 and a Ph.D. degree in technical sciences in 2002. His doctoral thesis addressed the sensor fusion problem in time-triggered systems. Wilfried Elmenreich has contributed significantly to the development of the TTP/A field-bus protocol and the standardization of the OMG Smart Transducer Interface Standard. In the last five years, Wilfried Elmenreich has published over 40 papers in the field of embedded real-time systems.

Click here for lecturer's website.