Special sessions
Papers are solicited for the following special sessions to be held during ICDSC 2009.
- Note that paper submission and review schedule for special session papers is the same as for regular papers.
- To submit a paper to a special session, please select the appropriate category in the conference submission server.
Special session 1:
Embedded techniques for smart cameras
Organisers: R. Kleihorst (VITO and U.Gent, Belgium), F. Berry (U. Blaise-Pascal, France)
Call for papers.
To achieve a real-time working smart camera system, effort is needed to port methods and algorithms to a computational system that is able to perform the task in real-time. Often, this entails interfacing to sensors and actuators, networking and mapping to dedicated hardware platforms, FPGAs or special processors. The immense engineering, innovation and research challenges of this part of the work is often overlooked or waved away, whereas especially this part makes things fly.
In this special session on embedded techniques we solicit papers on topics such as:
- mapping of algorithms on dedicated platforms
- interfacing and networking
- smart camera vision processor design
Special session 2:
Activity monitoring by multi-camera systems
Organiser: Tiziana D´Orazio (SSIA-CNR, Italy)
Call for papers.
Activity monitoring from video is one of the most promising applications of computer vision. In recent years, this problem has caught the attention of researchers from industry, academia, security agencies, consumer agencies, and also the general public. With the proliferation of inexpensive cameras and the deployment of high-speed, wireless networks, it has become economically and technically feasible to use multiple cameras for activity monitoring. By analyzing information from more cameras it is possible to monitor many large, complex areas such as buildings, airports, road and rail networks, sport grounds. Besides multi-camera systems can provide more confident deductions about the events of interest and reduce ambiguities in a single view caused by occlusion or other factors.
This special session serves as a forum for researchers working on multi-camera systems to discuss current challenges and to present the latest research results.Research in this field addresses large areas such as:
- visual cue integration from multi-camera observations
- AI-based scene analysis and learning
- spatial temporal object interactions for event recognition
- behavior analysis
- multi-camera calibration
- scene reconstruction.
Special session 3:
Advances in coding for distributed camera networks
Organisers: C. Abhayaratne (Univ. of Sheffield, UK), M. Mrak (Univ. of Surrey, UK), F. Dufaux (EPFL, Switzerland)
Call for papers.
Recent advances in sensor networks, distributed processing and computational vision have led to the emergence of context aware and user centric smart distributed camera and vision applications, such as, distributed and immersive multimedia, surveillance, smart-environments and multi-view vision. In most of these applications, cameras are operated in low power and low resources environments and the captured video is transmitted via error-prone wireless links. These conditions demand low complex highly robust video compression schemes.
The latest video compression paradigms focus on reducing coding/decoding complexity, enabling high robustness to error-prone channel transmissions, exploiting inter-view redundancies for high coding gains, ability to dynamic reconfiguration of decoding and incorporating features like scalability and random accessibility, to support fast and efficient compressed domain information extraction and processing. In particular, recent years have seen efforts in scalable video coding, distributed video coding, multi-view video coding and reconfigurable video coding in video coding research and standardisation. In addition to coding technology, MPEG-21, the Multimedia interface standard, provides means of describing and adapting content bit streams in an effective and fast manner, so that analysis and visualisation of content is possible formats, spatial-temporal resolutions, visual quality, and resource requirements suitable for the target applications. Recent years have also seen an increased tendency towards analysis of visual content on compressed domains, with only partial decoding of bit streams.
The aim of this special session is to present the influence of these emerging compression schemes and compressed domain analysis technologies in addressing the current challenges in distributed camera networks. We hope to bring together the two fields of content coding and content analysis with respect to solving distributed camera challenges and create a forum for interdisciplinary applications of video coding research. Contributions are solicited in the broad areas of compression and compressed domain analysis for distributed camera networks. Topics include, but are not limited to:
- distributed video coding
- multi-view video coding
- scalable video coding
- reconfigurable and distributed video coding architectures
- error resilience and concealment
- low complexity encoders
- content adaptation
- compressed domain analysis
Special session 4:
Collaborative Signal Processing for Distributed Systems
Organiser: D. Mandic (Imperial College, UK)
Call for papers.
Research in smart camera networks involves a number of related challenges from different disciplines. One of them is to deal with information at the signal level and try to use the information from a number of heterogeneous or low resolution sensors as efficiently as possible. The focus of this session is on the algorithms and solutions from the signal processing community, which may be of interest to the research community working on smart cameras.
This session will help reveal the extent to which applications of smart camera systems can benefit from intelligent pre-processing at the sensor level and will bring together latest research and practical solutions in collaborative and distributed signal processing for real world systems. This includes:
- low level conditioning of multidimensional signals
- shade and illuminant removal
- algorithms for input variable selection
- image fusion, dealing with multidimensional data natures
- fusion of information from heterogenous sensors, e.g. video and audio
- conditioning of video streams
Special session 5:
Dependability, Security and Privacy in Smart Camera Systems
Organizer: Dimitrios Serpanos (Industrial Systems Institute & University of Patras, Greece)
Call for papers.
Distributed smart camera systems are becoming increasingly important, being used in public and private environments for a range of surveillance applications related to security and safety. The complexity of the systems is continuously increasing, considering the complexity of the smart cameras themselves as well as the complexity of their interconnection and the related applications.
The applications of smart camera systems typically involve the collection, storage and processing of sensitive and/or personal information, leading to strong system requirements for security and privacy. Furthermore, the criticality of several applications leads to strong dependability requirements at the system level, achieving continuous operation and increased fault tolerance in light of accidental or malicious faults. As the technological area of distributed smart camera systems is interdisciplinary, combining advanced computing -embedded or not-, networking, sensing and signal processing technologies, the requirements for security, privacy and dependability lead to sophisticated solutions, combining security and dependability of nodes (e.g., cameras), networks and applications, taking into account the resource restrictions of embedded systems.
This special session focuses on technologies for dependability, security and privacy in smart camera systems. Topics of interest include, but are not limited to:
- secure and dependable smart camera systems (technologies and architectures)
- embedded systems security with applications to smart cameras
- sensor network security and dependability
- surveillance application security and privacy technologies
- security and privacy in video systems and services
Further Information: bernhard.rinner@uni-klu.ac.at