Hive is a novel middleware system for distributed vision computation, providing mechanisms for simple setup and configuration of vision processing modules. Hive facilitates communication between independent cross-platform modules via an extensible protocol, allowing arbitrary connection of modules to form a processing pipeline.

A plug-in interface allows general software to be represented as Hive modules: e.g. drivers for hardware devices such as cameras or implementations of particular vision algorithms. The modules are set up as a peer-to-peer network which allows for automated data transfer, callbacks and synchronisation. The architecture, communication protocol, plug-in interface and control system for the modules are all described, and a proof of concept distributed face tracking system is presented to demonstrate the simplicity and flexibility for creating complex distributed vision applications using Hive.

Presented in Stanford University, California, September 2008 at the International Conference on Distributed Smart Cameras.