Implementations of deep learning models applied on super-resolution for Deep Learning for Computer Vision Project (EECS 598). Single image super-resolution aims to recover a high-resolution image from a single low resolution image, is a classical problem in computer vision.
This project contains PyTorch implementations of different architectures for single image super-resolution. The implemented networks include:
-
The super resolution convolutional neural network described in Image Super-Resolution Using Convolutional Networks (C. Dong, et. al. 2014)
-
The Fast-super resolution convolutional neural network described in Accelerating the Super-Resolution Convolutional Neural Network