Neural Image Super-Resolution¶
Image Super-Resolution¶
Implementation of Image Super Resolution CNNs in Keras from the paper Image Super-Resolution Using Deep Convolutional Networks.
Also contains a modular framework, which allows a variety of other super resolution models to be trained and distilled :
1) Denoising Autoencoder SR models
2) ResNet SR models
3) Efficient SubPixel Convolutional SR models
4) Distilled ResNet SR models
5) Non-Local ResNet SR models (experimental)
Image Super-Resolution using GANs¶
An incomplete project that attempts to implement the SRGAN model proposed in the paper Photo-Realistic Single
Image Super-Resolution Using a Generative Adversarial Network in Keras.