Workshop I: Workshop on Smart Image Sensor Systems
- Prof. Marilyn Wolf, Georgia Institute of Technology, [email protected]
- Prof. Ricardo Carmona Galan, IMSE-CNM, [email protected]
Smart image sensor systems that integrate sensing, computation, and memory are the next generation of smart cameras. Traditional smart cameras combine traditional sensors and embedded computing systems-on-chips to perform in-camera analysis. Smart sensor systems may combine analog and digital computation with the sensor, resulting in significantly lower latency computations as well as smaller SWaP (size, weight, and power). The novel architectures presented by smart image sensors call for a new round of algorithm/architecture co-design to leverage the capabilities of these systems. Potential topics include:
- Analog and mixed-signal on-sensor computation.
- Digital on-sensor computation.
- 3D VLSI architectures for smart sensors.
- Computer vision applications (tracking, target identification, etc.) adapted to a chosen smart sensor platform.
- Algorithm-architecture co-design for smart sensor systems.
- Simulation studies of smart sensor systems.
- Applications and use cases for smart sensor systems.
- Distributed networks of smart sensors.
- Tutorials on smart image sensors.
Workshop II: Human Behavior Analysis based on Video Analytics and Deep Learning
- Dr. Maheshkumar H Kolekar, Inidian Institute of Technology Patna, [email protected]
- Dr. R. Balasubramanian, Indian Institute of Technology Roorkee, [email protected]
Human behavior recognition (HBR) is a key area in video analysis and machine learning application. Close circuit cameras are now widely used to increase security in many high alert areas such as airport, bus stop and railway station. There are various methods for analyzing human behavior such as sensor based activity recognition, gait analysis based activity recognition. Since video is a low cost and easy to use device for monitoring activities, video based human behavior analysis is widely used. Video analytics have application in patient’s behavior analysis also. Deep learning allows models that are composed of multiple processing layers to learn features of data with multiple levels of abstraction. Now-a-days deep learning techniques such as Convolutional Neural Network, Recurrent Neural Network have become popular for human activity recognition. The workshop aims at bringing computer vision and machine learning experts together for presenting their novel ideas in HBR using video processing and deep learning approach. This will give a platform to researchers from both fields to discuss their original ideas in HBR, solution of existing HBR problems and look for future scope in designing more robust model for human behavior analysis system. Scope and topics of interest include, but are not limited to:
- Human behavior analysis
- Human activity recognition
- Human/ Object tracking and classification
- Crowd behavior analysis
- Deep learning techniques application in human activity recognition
- Video analytics for patient behavior analysis
Workshop III: Multi-source Surveillance for Public Safety and Healthcare
- Dr. Mukesh Saini, Indian Institute of Technology Ropar, [email protected]
- Dr. Dwarikanath Mahapatra, IBM Research Australia, [email protected]
With rapid growth in sensing and communication technology, we have various new sources of information that can complement video to achieve robust and reliable surveillance, e.g. heat detector, smoke detector, infrared/thermal sensor, motion sensor, metal detectors, GPS, weather sensors, and various other sensors found on current smartphones. Additionally, online social networks and real-time news portals provide valuable information about individuals and surrounding environment. We can assimilate all these data from different sources to make accurate and reliable situation assessment using artificial intelligence (AI) and Deep learning (DL) methods. This workshop aims to bring together researchers that are specifically working on applying advanced learning techniques to exploit multi-source information for public safety and healthcare. The workshop aims to build a common platform for synergizing research efforts and develop innovative methods population health and safety. We solicit original and unpublished research papers that analyse multiple source for safety and healthcare applications. Topics for the workshop include, but are not limited to:
- Multi-source data analysis for security threat and crime detection
- Multi-source data analysis for forensics
- Multi-source data analysis for abnormal situation monitoring
- Advanced learning techniques for multi-source data analysis for safety and health
- Multi-source data analysis for population health management
- Multi-source data analysis for disease detection, diagnosis and treatment plans
- Online social network analysis for safety and health monitoring
- Multi-source data analysis for remote or tele-health management
Researchers working on multimodal analysis of single source (such as image, audio, text) for public safety and healthcare are also encouraged to submit their original work. After the workshop, we intend to submit a special issue proposal in a reputable journal for extended papers of the workshop on the same topic.
Workshop IV: Security of Video Surveillance Systems (SVSS 2018)
- Prof. Yuval Elovici, Ben-Gurion University of the Negev, [email protected]
- Dr. Asaf Shabtai, Ben-Gurion University of the Negev, [email protected]
- Mr. Oleg Brodt, Ben-Gurion University, [email protected]
Video surveillance systems have become an important part of our lives, introducing new capabilities to improve civilian protection, reduce crime, increase the security of military and critical infrastructures, etc. As the trend of the Internet of Things (IoT) has evolved, video and signaling surveillance systems have begun to integrate new, advanced, and intelligent functionalities, such as real-time video stream analysis and detection using advanced machine learning techniques, streaming data to servers deployed in the cloud to provide access to the content by multiple devices from any location, and more. However, the integration of new functionalities into these systems exposes them to new risks that need to be considered and mitigated. For example, we cannot assume that the transmitted and visualized video is protected and reflects the true state of the monitored environment/objects, since an attacker can monitor and manipulate the data streams of such systems in real-time (e.g. reply attack) from a remote location. Some of the security problems resulting from the intersection of the digital and physical realms, introduce cyber security challenges which are unique to video and signaling surveillance systems and may cost human lives. It is therefore critical to address the emerging cyber security related problems, which can affect (compromise) various components of such surveillance systems. This workshop aims to bring together the latest cyber security research pertaining to all aspects of video surveillance systems. We are soliciting contributions on (but not limited to) the following topics:
- Secured architecture for video surveillance systems
- Intrusion detection in video surveillance systems
- Inferring information from encrypted video streams
- Cyber attacks on video surveillance systems
- Cyber attacks from video surveillance systems
- Adversarial AI aimed at video streams and surveillance systems
- Surveillance system honeypots
- Security measures for surveillance systems
Workshop V: Traffic and Street Surveillance for Safety and Security (T4S)
- Marco Del Coco, Institute of Applied Sciences and Intelligent Systems (CNR, Italy), [email protected]
- Siwei Lyu, University at Albany, [email protected]
- Pierluigi Calcagnì, Institute of Applied Sciences and Intelligent Systems (CNR, Italy), [email protected]
- Ming-Ching Chang, University At Albany, [email protected]
In the last couple of decades, there has been a tremendous increase in demand for smart systems capable of monitoring street security and traffic states. Big city centers are subjected to a growing amount of people and vehicle mobility, making control and security management increasingly difficult. The contribution of automatic systems capable to centralize street control and automatically detect possible problems (accidents, traffic jam, brawls, etc.) in order to elicit and support the intervention of law enforcement agencies or medical staff is needed. The proposed workshop is organized jointly with the Challenge on Advanced Traffic Monitoring, and it is addressed to researchers interested in contributing on the general topic of security on mobility and street applications. Research papers are solicited in, but not limited to, the follow areas and topics:
- Vehicle detection and tracking
- Vehicle type classification
- Pedestrian detection and tracking
- Vehicle/pedestrian behaviour
- Traffic jam detection
- Car/pedestrian accidents
- Activity monitoring systems
- Scene understanding
- Visual attention and visual saliency
- Matching vehicles across cameras
- Smart environments
- Safety and security
- Technology for cognition
- Navigation systems
- Sensory substitution
- Datasets and evaluation procedures
Workshop papers are 6 pages long (including references), with up to two more pages (attracting extra page charges), with double blind review. Submission: 7 September 2018
Notification of acceptance: 1 October 2018
Camera-ready deadline: 15 October 2018