Operational significance of receptive field assemblies (Q1106144)

The author describes how 2-dimensional differential geometry may be implemented in a way that, he claims, closely resembles the known structure of the visual system. He uses the tangent bundle, the cotangent bundle, sometimes conformal Riemannian structure, then again Riemannian structure without clear distinction, connections and jet bundles of the fourth order (he claims biological evidence of this). Surprisingly curvature does not play a role. All entities are blurred with some Gaussian integrals. The paper is vague and sketchy.