Ethan Harris (@ethanwharris), Mahesan Niranjan and Jonathon Hare (@jonhare)
This repository contains the code for our paper 'Foveated Convolutions: Improving Spatial Transformer Networks by Modelling the Retina', accepted to the NeurIPS 2019 workshop on Shared Visual Representations in Humans and Machines (SVRHM).
The main file for running experiments is main.py
, uncomment the relevant sections to run different configurations.
ResNet experiments and the code to generate the figure can be found in train_resnet.py
and figure.py
respectively.