Self-Feature-Super-Resolution

IMAGE SUPER-RESOLUTION USING CNN OPTIMISED BY SELF-FEATURE LOSS

Pre-trained model and testing codes (PyTorch)

Overview:

This script generates 3x super resolved images based on the test images in the folder "test_images/". The model is trained from the project of the "Image Super-Resolution Using CNN with Self-Feature Loss".

Network Architecture:

Results:

Installation Dependencies:

• Pytorch
• PIL
• Numpy
• Torchvision

Citation

Zhao Gao, Eran Edirisinghe, Slava Chesnokov

Image Super-Resolution Using CNN Optimised by Self-Feature Loss

*Submitted to ICIP 2019

@inproceedings{gao2019image,

title={Image Super-Resolution Using CNN Optimised By Self-Feature Loss},

author={Gao, Zhao and Edirisinghe, Eran and Chesnokov, Slava},

booktitle={2019 IEEE International Conference on Image Processing (ICIP)},

pages={2816--2820},

year={2019},

organization={IEEE}

}