An imaging architecture based on derivative estimation sensors

Abstract

An imaging architecture is proposed, where the first and second derivatives of the image are directly computed from the scene. Such an architecture bypasses the problems of estimating derivatives from sampled and digitised data. It, therefore, allows one to perform more accurate image processing and create more detailed image representations than conventional imaging. This paper examines the feasibility of such an architecture from the hardware point of view. © 2009 Springer-Verlag Berlin Heidelberg.