Multi-camera systems have attracted attention in recent years due to rapidly dropping cost of digital cameras. This has enabled wide variety of new research topics and applications for Multi View Imaging (MVI) systems. Virtual view synthesis, high performance imaging, image/video segmentation, depth map estimation, 3DTV, and free viewpoint TV are examples of such applications. High data rates, camera synchronization and calibration make the real-time deployment of these systems a challenging task. Most built multi-camera systems were designed as multi-video recording systems, where their target applications were processed offline on computers. In a centralized processing approach, all the computation of a multi-camera system takes place on a single central processing unit (e.g. PC). This approach does not enable real-time deployment if the number of cameras is high. The processing demand and connectivity problem from the cameras to a single unit, are limiting factors in this approach. One solution for real-time deployment is to distribute and run in parallel the target applications at the camera level and among the cameras. This method assumes intra (at the camera level) and inter (among the camera) signal processing for the cameras. In the inter signal processing, exchange of information (image/data) takes place among the cameras. Hence a communication medium is required for inter signal processing. This thesis introduces a communication medium for the support of inter signal processing in a multi-camera system based on the interconnected network concept. A methodology is introduced for network topology selection, camera assignments to network nodes, performance analysis and simulation of the network for any target application.