Tony Vladusich (@[email protected])

Attached: 1 image Every single model of human vision is wrong. Not just a bit wrong, I mean just flat out wrong from the get go. The reason is that every single model incorrectly assumes that human vision is space invariant and passive, despite the obvious space variant nature of the retinal sensory array and the constant movements of the eyes to foveate targets. (Why these models incorporate false assumptions about the nature of vision continues to baffle me.) Yet, almost nobody knows about the incredible seminal work of the late Eric L Schwartz. (Eric was my friend and mentor in my years at Boston University.) If anybody deserved a Nobel prize for applying physics principles and techniques to neuroscience, it was him. Eric died in 2019, so he won't be winning any prizes, let alone Nobel prizes. (He'd be laughing in his grave at the fact that Hinton and Hopfield won the prize, and his long time colleague and "rival" Stephen Grossberg was snubbed.) Eric applied methods of complex analysis, vector calculus, and conformal mapping to create a rigorous framework for understanding the computational architecture and functional properties of the primate visual cortex. His work focused on space variant active vision and computational neuroscience (in fact, he coined both terms). Here is one of his great review papers on the computer vision applications of his work. https://www.sciencedirect.com/science/article/pii/0893608095000925