A demonstration of Deep Network Interpolation for SR application. I use pretrained SRGAN/ESRGAN provided by MMSR.
python main.py --input path/to/your/image
The interpolation result on Set14
baboon is here!
The Perception-Distortion Tradeoff was observed.
- Wang, Xintao, et al. "Deep network interpolation for continuous imagery effect transition." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2019.
- MMSR from Multimedia Laboratory, CUHK.- https://github.com/open-mmlab/mmsr
- Bevilacqua, Marco, et al. "Low-complexity single-image super-resolution based on nonnegative neighbor embedding." (2012): 135-1.
- Blau, Yochai, and Tomer Michaeli. "The perception-distortion tradeoff." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018.