Hamid Aghajan, Andrea Cavallaro
Distributed vision networks is a multi-disciplinary research field that defines rich conceptual and algorithmic opportunities for the fields of computer vision, signal processing, pervasive computing, and wireless sensor networks. It also creates opportunities for design paradigm shifts in these fields given its emphasis on distributed and collaborative fusion of visual information, enabling researchers in the named areas to participate in the creation of novel smart environment applications that are interpretive, context aware, and user-centric in nature.
Technological advances in the design of image sensors and embedded processors have facilitated the development of efficient embedded vision-based techniques. Design of scalable, network-based applications based on high-bandwidth data such as images requires a change of paradigm in the processing methodologies. Instead of streaming raw images to a powerful, central processing unit, in the new paradigm each network node utilizes local processing to translate the observed data into features and attributes, which are shared with other network nodes to make a collaborative deduction about the event of interest. Building upon the premise of distributed vision-based sensing and processing, many novel application areas in smart environments such as patient and elderly care, ambient intelligence, multimedia and gaming can be enabled.
Design based on access to a networked set of sources of visual information provides researchers in the vision processing discipline with novel research opportunities not only through introducing the notion of spatial collaboration in data fusion among the network nodes, but also in considering the effective ways of network-wide collaboration and levels in which data can be exchanged between the nodes. In such a framework data fusion can occur across the three dimensions of 3D space (multiple views), time, and feature levels. Joint estimation and decision making techniques need to be developed taking into account the processing capabilities of the nodes as well as the network bandwidth limits and application latency requirements. Spatiotemporal data fusion algorithms employing information obtained by the network across the dimensions of space, time, and feature levels and the impacts of the various cost and efficiency tradeoffs need to be examined against the requirements of the application. In addition, deductions produced by collaborative processing in the network can be used as feedback to each camera node to enable active vision techniques by instructing each camera as to which features may be of importance to extract in its view.
The target audience for the course consists of researchers active in various signal processing applications for camera networks applications such as human presence and gesture analysis, as well as graduate students involved in vision algorithm design research. The course offers a perspective of the various methodologies based on the flexibilities and tradeoffs introduced by distributed vision sensing and processing. As a result of providing such perspective, the course aims to encourage participation of vision researchers in developing novel algorithms based on the potentials of distributed camera networks.
Offering insight into the potentials and challenges of distributed vision networks and the novel design methodologies employed by leading research groups working in this area is the objective of the proposed short course. By emphasizing the distributed processing aspects of multi-camera networks, the proposed course inspires participation in this field by algorithm design researchers, and promotes the opportunity of making an impact to the field from a signal processing perspective, a timely and potentially rewarding proposition for the ICASSP community.
Hamid Aghajan is an associate professor of Electrical Engineering (consulting) at Stanford University, where he supervises the Wireless Sensor Networks Laboratory (WSNL). Hamid's research is focused on distributed vision networks. Prior to joining Stanford in 2003, Hamid had 9 years of experience in technical and managerial positions in high-technology companies leading product development in image processing, wireless communications, and semiconductor manufacturing systems. During his industrial experience, Hamid was a co-founder and vice president of a start-up company developing optical filters for fiber telecommunications. Recent work in Hamid's research group consists of vision-based algorithms for assisted living, occupancy sensing for smart buildings, multimedia and gaming based on human gesture, face, body model, and hand gesture analysis, and network localization - all based on distributed vision processing in multi-camera networks. He has organized and taught new courses on wireless sensor networks and vision networks at Stanford. Hamid is co-organizer and technical co-chair of the first International Conference on Distributed Smart Cameras (ICDSC 2007), and general co-chair of ICDSC 2008, organizer of short courses on Distributed Vision Processing in Smart Camera Networks at CVPR 2007, 2008 and ACIVS 2007, organizer of a special session on Distributed Processing in Image Sensor Networks at ICASSP 2007, guest editor of the IEEE J-STSP special issue on Distributed Processing in Vision Networks, organizer of a special session on Distributed Processing in Image Sensor Networks at ICASSP 2007, co-chair of the Behaviour Monitoring and Interpretation (BMI) workshop at German AI Conference 2008, special session chair of Vision-based Reasoning at AITAmI workshop at ECAI 2008, workshop co-chair of Multi-camera and Multi-modal Sensor Fusion Algorithms and Applications at ECCV 2008, special session chair of Multi-Sensor HCI for Smart Environments at Face and Gesture Conference 2008, and workshop co-chair of Vision Networks for Behaviour Analysis (VNBA) at ACM Multimedia 2008.. He has published numerous journal and conference papers and holds 5 US patents. Hamid obtained his Ph.D. degree in Electrical Engineering from Stanford University in 1995.
Andrea Cavallaro is Reader (Associate Professor) in Multimedia Signal Processing at the Department of Electronic Engineering, Queen Mary, University of London (QMUL). 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), workshop co-chair of Multi-camera and Multi-modal Sensor Fusion Algorithms and Applications at ECCV 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 and of the Special Issue of Multi-camera and multi-modal sensor fusion’, Computer Vision and Image Understanding. He 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 ''Hybrid particle filter and mean shift tracker with adaptive transition model'' and “Particle PHD filtering for multi-target visual tracking”, winner of the student paper contest at the IEEE ICASSP in 2005 and 2007, respectively. He has been a member of the organizing/technical committee of several conferences, such as IEEE ICME, IEEE ICIP, SPIE VCIP, ACM Multimedia, IEEE AVSS, ACM/IEEE ICDSCECCV-VS, PETS. He acts as reviewer for several international conferences and journals, and he is author of more than 70 papers, including 5 book chapters.