This project is implemented by Caffe, Python and Matlab RSRCNN_train: resudial image super-resolution convolutional neural network is used to train the model for fine-tuning the RACNN.
Classification_CNNs: it contains two networks Alexnet and VGGNet, and Python codes (notebook) are provided to train the RACNN.