Deep Video Super-Resolution Network Using Dynamic Upsampling Filters Without Explicit Motion Compensation

This is a tensorflow implementation of the paper. PDF

directory

./inputs/G/ Ground-truth video frames
./inputs/L/ Low-resolution video frames

./results/<L>L/G/ Outputs from given ground-truth video frames using depth network
./results/<L>L/L/ Outputs from given low-resolution video frames using depth network

test

Put your video frames to the input directory and run test.py with arguments <R>, <L> and <T>.

python test.py <R> <L> <T>

<R> is the upscaling factor of 2, 3, 4. <L> is the depth of network of 16, 28, 52. <T> is the type of input frames, G denotes GT inputs and L denotes LR inputs.

For example, python test.py 4 16 G super-resolve input frames in ./inputs/G/* using 16 depth network with upscaling factor 4. (Possible combinations for <R> <L> is 2 16, 3 16, 4 16, 4 28, and 4 52.)

This code was tested under Python 2.7 and TensorFlow 1.3.0.

video

bibtex

@InProceedings{Jo_2018_CVPR,
	author = {Jo, Younghyun and Oh, Seoung Wug and Kang, Jaeyeon and Kim, Seon Joo},
	title = {Deep Video Super-Resolution Network Using Dynamic Upsampling Filters Without Explicit Motion Compensation},
	booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
	year = {2018}
}