Pytorch implementation of HighRes-net, a neural network for multi-frame super-resolution, trained and tested on the European Space Agency’s Kelvin competition. This is a ServiceNow Research project that was started at Element AI.

DeepSUM: Deep neural network for Super-resolution of Unregistered Multitemporal images (ESA PROBA-V challenge)

Permutation invariance and uncertainty in multitemporal image super-resolution

Enhance low resolution image using Convolutional Neural Nets. My solution to European Space Agency's one of its many contest.

Project to test the capabilities of deep learning architectures for the super-resolution task using hyperspectral images published by the PROBA-V satellite [old].