OPENCV SUPERRES

SuperResolution in OpenCV.

Requirements

• OpenCV4

Build

Type:

make

Usage

opencv_superres image_in fsrcnn 2 models/FSRCNN/models/FSRCNN_x2.pb image_out.png

Models

• EDSR
• ESPCN
• FSRCNN
• LapSRN

TensorFlow EDSR

TensorFlow implementation of Enhanced Deep Residual Networks for Single Image Super-Resolution[1].

It was trained on the Div2K dataset - Train Data (HR images).

References

[1] Bee Lim, Sanghyun Son, Heewon Kim, Seungjun Nah, and Kyoung Mu Lee, "Enhanced Deep Residual Networks for Single Image Super-Resolution," 2nd NTIRE: New Trends in Image Restoration and Enhancement workshop and challenge on image super-resolution in conjunction with CVPR 2017. [PDF] [arXiv] [Slide]

TensorFlow ESPCN

Tensorflow implementation of ESPCN algorithm described in [1]. This project was done during the Google Summer of Code 2019 program with OpenCV [2].

References

[1] Shi, W., Caballero, J., Huszár, F., Totz, J., Aitken, A., Bishop, R., Rueckert, D. and Wang, Z. (2019). Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network. Available at: https://arxiv.org/abs/1609.05158
[2] https://summerofcode.withgoogle.com/projects/#4689224954019840
[3] Agustsson, E., Timofte, R. (2017). NTIRE 2017 Challenge on Single Image Super-Resolution: Dataset and Study. Available at: http://www.vision.ee.ethz.ch/~timofter/publications/Agustsson-CVPRW-2017.pdf

https://data.vision.ee.ethz.ch/cvl/DIV2K/

TensorFlow FSRCNN

TensorFlow implementation of Accelerating the Super-Resolution Convolutional Neural Network [1].

This implementation replaces the transpose conv2d layer by a sub-pixel layer [2].

Includes pretrained models for scales x2, x3 and x4. Which were trained on T91-image dataset, and finetuned on General100 dataset.

FSRCNN-small is a network with fewer parameters. Thus it is faster but has lower performance.

References

1. Chao Dong, Chen Change Loy, Xiaoou Tang. Accelerating the Super-Resolution Convolutional Neural Network, in Proceedings of European Conference on Computer Vision (ECCV), 2016
2. Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network. By Shi et. al.

TensorFlow LapSRN

Tensorflow implementation of LapSRN algorithm described in [1]. It can now support training for 2x, 4x, and 8x scaling factor.

References

[1] Lai, W., Huang, J., Ahuja, N. and Yang, M. (2019). Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks. Available at: https://arxiv.org/abs/1710.01992
[2] Agustsson, E., Timofte, R. (2017). NTIRE 2017 Challenge on Single Image Super-Resolution: Dataset and Study. Available at: http://www.vision.ee.ethz.ch/~timofter/publications/Agustsson-CVPRW-2017.pdf

https://data.vision.ee.ethz.ch/cvl/DIV2K/

Protocol Buffers

Google's data interchange format.