This is the code for our ECCV20 Paper DGDMLSR [WIP]

Requirement:

Python 3.7 with cuda10.1 is needed

install these python packages:

torch 1.2 torchvision scikit-image tqdm opencv-python

Data generation

Use example image

Step.1 use get_image.py to collect NYU depth images and depth maps.
Step.2 use depth_inv.py and gen_patch.py to collect patches from the image and its depth map.
Step.3 use gen_pkl.py to generate python pickle file to accelerate the dataloader.

Use your own image

use a depth estimation model to obtain the depth map.
use inverse_depth to get the inversed depth
follow the example

Training

We train a specific model for each image.

use main.py to start training, you will get the checkpoints and super resolved images for each epoch.

Testing

You can use the scoring script from PIRM18 to test the NIQE and PI score. You need matlab >= R2016 to process these code.