Demo abstract: Fusion of audio and image information for efficient object detection and capture

This abstract describes the demonstration of a tightly integrated audio-image wireless multimedia sensor network (WMSN) for object detection and capture. By making use of available multi-modal information, we are able to quickly and efficiently localize objects using relatively cheap audio and camera sensors with more success than would be possible based only on one modality alone. While similar, more complex, techniques have been considered before the novelty of the approach is in its practical implementation on a relatively low-cost, resource-constrained WMSN.