https://github.com/sovit-123/deep-learning-image-super-resolution

This is a deep learning project based on the Image Super-Resolution Using Deep Convolutional Networks - SRCNN paper using the PyTorch deep learning library.
https://github.com/sovit-123/deep-learning-image-super-resolution

computer convolutional-neural-networks deep-learning image-processing image-super-resolution neural-network python pytorch pytorch-implementation

Last synced: about 2 months ago
JSON representation

This is a deep learning project based on the Image Super-Resolution Using Deep Convolutional Networks - SRCNN paper using the PyTorch deep learning library.

• Host: GitHub
• URL: https://github.com/sovit-123/deep-learning-image-super-resolution
• Owner: sovit-123
• Created: 2020-06-19T05:25:54.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2020-06-24T13:42:14.000Z (almost 6 years ago)
• Last Synced: 2025-03-29T00:42:30.044Z (about 1 year ago)
• Topics: computer, convolutional-neural-networks, deep-learning, image-processing, image-super-resolution, neural-network, python, pytorch, pytorch-implementation
• Language: Python
• Homepage:
• Size: 23.5 MB
• Stars: 3
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Deep Learning Image Super Resolution

## What is the Project About

* This is a deep learning project based on the [Image Super-Resolution Using Deep Convolutional Networks - SRCNN](http://mmlab.ie.cuhk.edu.hk/projects/SRCNN.html) paper using the **PyTorch** deep learning library.

## Framework and Dependencies

* **The project is built on PyTorch 1.4.** 

* You will need MATLAB to execute the `.m` files.

## Directory Structure

* The following is the directory structure to arrange everything for the project.

  ```

  ├───input

  │   ├───bicubic_2x

  │   ├───bicubic_4x

  │   ├───bicubic_rgb_2x

  │   ├───bicubic_rgb_4x

  │   ├───General100

  │   ├───Set14

  │   ├───Set5

  │   ├───T91

  │   ├───T91_G100

  |    train_mscale.h5

  ├───outputs

  └───src

  ```

* `input`: contains the datasets that are used for training and testing. The `train_mscale.h5` is the training datasets that gets generated after running the `generate_train.m` file. 

  * Currently the model has been trained on both `T91` and `General100` image datasets. Both of these datasets are merged into `T91_G100` folder. The same corresponds to in the `generate_train.m` file.

  * The `bicubic_x` folders contain the blurred images that we use for testing. Generate those images using the `bicubic.py` file inside the `src` folder.

* The `outputs` folder will contain all the output files along with the trained model.

* `src` contains the python and MATLAB files.

***Note***: *I have take the MATLAB codes from the [SRCNN-Keras](https://github.com/YapengTian/SRCNN-Keras) repository. The original `generate_train.m` file generate greyscale sub-images. I have formatted the code so as to generate colored (RGB) sub-images. As such, in this project, you will be able to train a neural network model that can carry out super-resolution on RGB images.*  *Please go through the code for more details*.

## Dataset

* You will find the datasets used in this project and more super-resolution datasets [here](https://github.com/xinntao/BasicSR/wiki/Prepare-datasets-in-LMDB-format).

## Execution

* `generate_train.m`: To generate the `train_mscale.h5` sub-images.

* Execute the python scripts while being within the `src` folder in the terminal.

  * `python bicubic.py --path ../input/Set14 --scale-factor 2x`: To create low-resolution bicubic images for the Set14 data by a scaling factor of 2x.

  * `python train.py`: For training the SRCNN model.

  * `python test.py --input ../input/bicubic_rgb_2x`: To generate high resolution images of the 2x scaled low-resolution images.

## Some Results

* The following are from testing on the 2x scaled low-resolution images.

![](https://github.com/sovit-123/Deep-Learning-Image-Super-Resolution/blob/master/input/bicubic_rgb_2x/baboon.png?raw=true)

![](https://github.com/sovit-123/Deep-Learning-Image-Super-Resolution/blob/master/outputs/outputs_baboon.png?raw=true)

![](https://github.com/sovit-123/Deep-Learning-Image-Super-Resolution/blob/master/input/bicubic_rgb_2x/comic.png?raw=true)

![](https://github.com/sovit-123/Deep-Learning-Image-Super-Resolution/blob/master/outputs/outputs_comic.png?raw=true)

![](https://github.com/sovit-123/Deep-Learning-Image-Super-Resolution/blob/master/input/bicubic_rgb_2x/zebra.png?raw=true)

![](https://github.com/sovit-123/Deep-Learning-Image-Super-Resolution/blob/master/outputs/outputs_zebra.png?raw=true)

## References

* [Image Super-Resolution Using Deep Convolutional Networks - SRCNN](http://mmlab.ie.cuhk.edu.hk/projects/SRCNN.html).

* [ SRCNN-Keras](https://github.com/YapengTian/SRCNN-Keras): For the `generate_train.m` file to create the `train_mscale.h5` training data.

* [SRCNN-Tensorflow](https://github.com/jinsuyoo/SRCNN-Tensorflow): For the test images.