A new imaging architecture and alternative interpretation of the structure the human retina

Abstract

An architecture is proposed where the layers of cells that lie in front of the photosensitive cells of the human retina are interpreted to be estimators of the first and second order derivatives of the brightness of the imagined scene. Such an architecture bypasses the problems of estimating derivatives from sampled and digitised data, as they are estimated directly from the scene. It also offers an explanation on why the photosensitive sensors of the human eye are placed at the back of the eye, behind four other layers of cellls.