Increasing Image Resolution with Neural Networks
This video quickly shows how to increase image resolution with neural networks.
The basic idea is: you train a neural network with a low resolution input and backpropagates with a higher resolution image as output.
The training was made from 16x16 RGB to 32x32 RGB using CIFAR-10 dataset. The trained neural network can then be used with images of any size. Interesting isn't it?
This video is part of the Conscious Artificial Intelligence Project.