Next Generation Smart Camera Networks

点击翻译按钮获取中文摘要

Camera networks have received much attention in recent years, in part due to their suitability in a wide range of applications. We also bear witness to a sudden increase in the availability of camera-equipped mobile devices. The immediate access to cameras creates new opportunities for designing the next generation of Smart Camera Networks (SCNs) capable of unattended operation and supporting new forms of collaborative sensing over expansive areas. We propose to develop scientific & theoretical foundations, investigate new methods, and research design principles needed to develop such camera networks.***Control and coordination in SCNs: We will develop scientific theory and mathematical machinery to study and characterize distributed sensing and control in ad hoc networks of non-stationary smart cameras, focusing in particular upon new models for collaborative sensing in such networks.***Networks of camera-equipped mobile devices: We will explore design principles and develop mathematical underpinnings for setting up short-duration, tasks-specific, ad hoc camera networks of transient and highly dynamic camera-equipped mobile devices in a manner that does not adversely affect the usability, battery-life, security, and privacy of these devices. ***Crowd tracking and scene understanding in SCNs: We will develop scientific and theoretical foundations for dense crowd tracking using ad hoc networks of non-stationary, smart cameras. In particular, we will explore "analysis via synthesis" approaches for dense crowd tracking in camera networks, where virtual crowd simulations exhibiting the statistical properties of the observed crowd will be used to refine crowd analysis and facilitate scene understanding.***I envision significant benefits and numerous applications of this research program. The proposed research program provides many opportunities for unique and interdisciplinary HQP training. Furthermore, the outcomes of this program will directly benefit both local & international camera technology, video analytics, and digital entertainment industries. E.g., the theory and algorithms developed here will lead to more capable camera-equipped consumer devices that will support new forms of participatory sensing and collaborative scene analysis. Scientific, algorithmic, and technical contributions of this program will increase our understanding of distributed, collaborative scene analysis in large-scale ad hoc networks of non-stationary cameras, benefitting the wider camera networks and computer vision communities.***This proposal aims to carryout highly original, fundamental scientific and technical research needed to realize the next generation of smart camera networks, and my existing work on camera networks and image analysis makes me an ideal candidate for pursuing the goals identified herein.**