Design and implementation of a cluster based smart camera array application framework

With the advance in computing, imaging and rendering technologies, applications taking advantage of multiple view dynamic scene information provided by a camera array have become more and more popular and practical. However, many such systems proposed in previous works are usually project-specific, and hence, there is a need for a general-purpose application framework to enable rapid prototyping and implementation. In this paper, we present a middleware-like framework which provides foundational components and modularized workflow services which are commonly required in different camera array applications. In particular, our framework is targeted to computer cluster based camera array applications, in which multiple computing nodes are network-connected and used to control the cameras and perform distributed computations. Software abstraction of inter-camera message/data communications is provided to enable event-driven workflow and data access location transparency, which in turn enables many computations involving multiple-camera to be parallelizable for better efficiency. In this paper, we demonstrate the effectiveness and performance of our proposed framework by illustrating two practical applications for 3D video and multiple camera based background/foreground segmentation. Our results demonstrate that the proposed framework is very promising and that it satisfies our initial requirements.