Underpinning more reliable and robust computer vision tasks is the need to better model background and clutter. This could lead to better sensors that integrate such modeling to yield a better (more useful) signal (higher signal to noise, lower bit rate, more appropriate for the application). Applications span the whole spectrum but the most obvious are surveillance and biometrics.

The significance is immediate from the rather wide application in computer vision  together with the recognition that poor performance when faced with changes in illumination, interference from multiple non-target objects and other clutter, etc.; is a significant, if not the main, impediment to acceptance of automated vision technology for surveillance and biometrics.

The national importance/benefit lies particularly in the areas of civil and military surveillance (e.g., crowd monitoring, seeking, recognizing (or at least characterizing) and tracking specific individuals etc.). Environmental monitoring and protection applications are another area of huge national benefit and significance.

1. Robotic Aircraft for Environmental Modelling
2. Biometrics and surveillance.