https://github.com/rageworx/libsrcnn

Super-Resolution imaging with Convolutional Neural Network library for G++, Non-OpenCV model.
https://github.com/rageworx/libsrcnn

c cpp gcc gpp imaging mingw-w64 no-opencv resolution super super-resolution

Last synced: 22 days ago
JSON representation

Super-Resolution imaging with Convolutional Neural Network library for G++, Non-OpenCV model.

• Host: GitHub
• URL: https://github.com/rageworx/libsrcnn
• Owner: rageworx
• License: lgpl-3.0
• Created: 2018-08-08T09:03:35.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2024-06-20T04:52:38.000Z (10 months ago)
• Last Synced: 2024-06-20T16:28:53.949Z (10 months ago)
• Topics: c, cpp, gcc, gpp, imaging, mingw-w64, no-opencv, resolution, super, super-resolution
• Language: C
• Homepage: http://rageworx.pe.kr/search/SRCNN
• Size: 1.29 MB
• Stars: 16
• Watchers: 8
• Forks: 1
• Open Issues: 6
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# libsrcnn

### Super-Resolution imaging with Convolutional Neural Network

A stand-alone library for Super-Resolution, Non-OpenCV model related in these projects:

* https://github.com/rageworx/SRCNN_OpenCV_GCC

* https://github.com/shuwang127/SRCNN_Cpp.

## Introduction

This is an open source project from original of this:

**SRCNN_Cpp** is a C++ Implementation of Image Super-Resolution using SRCNN which is proposed by Chao Dong in 2014.

 - If you want to find the details of SRCNN algorithm, please read the paper:  

   Chao Dong, Chen Change Loy, Kaiming He, Xiaoou Tang. Learning a Deep Convolutional Network for Image Super-Resolution, in Proceedings of European Conference on Computer Vision (ECCV), 2014

 - If you want to download the training code(caffe) or test code(Matlab) for SRCNN, please open your browse and visit http://mmlab.ie.cuhk.edu.hk/projects/SRCNN.html for more details.

 - And thank you very much for Chao's work in SRCNN.

## WIKI

* See [Wiki](https://github.com/rageworx/libsrcnn/wiki) page.

* Includes OpenMP support for macOS.

## Additional references

* Fast resizing methods for BiCubic filtering 

    * [Free Image Project](http://freeimage.sourceforge.net/)

	* [librawprocessor](https://github.com/rageworx/librawprocessor)

    * [fl_imgtk](https://github.com/rageworx/fl_imgtk)

 

## Features

* Faster about 400% or more than OpenCV GCC version of SRCNN, with OpenMP.

    - references :

    - commit/b340b885a58605f793aa000eebc7f96c19c8e9fe#commitcomment-103507343

	- commit/b340b885a58605f793aa000eebc7f96c19c8e9fe#commitcomment-103507802

* None-OpenCV, no OpenCV required !

* Compilation availed for almost of POSIX g++.

* Simply optimized for basic OpenMP.

* Works well even without OpenMP like macOS.

* Not support M$VC, sorry MS guyz.

## Sample images

* Original 100%

    ![IMG_0](Pictures/butterfly.png)

* Bicubic 150%

    ![IMG_1](Pictures/butterfly_bicubic.png)

* SRCNN 150%

    ![IMG_2](Pictures/butterfly_srcnn.png)

* SRCNN (Convolution Y channel) 150%

    ![IMG_3](Pictures/butterfly_srcnn_convolution.png)

## Supporting platforms

* Windows 32, 64 with MSYS2 + MinGW-W64

* Almost any Linux, x86_32, x86_64, arm, armhf, aarch64

* macOS ( clang, llvm )

## Latest Changes

### Verison 0.1.10.40

* Better speed, less memory usage by convolution I+II

* Regards to zvezdochiot@github

## Previous Changes

### Verison 0.1.9.35

* Fixed memory bug in float images from RGB case.

* header version flag fixed.

### Verison 0.1.9.34

* Fixed don't use color space scaling with bicubic filter.

* Now supporting alpha channel.

### Verison 0.1.8.30

* Precision step scaling bug fixed.

### Verison 0.1.8.28

* Precision step scaling option availed.

* included option by reason of libsrcnn trained for maximum double multiply.

### Verison 0.1.6.23

* Fixed a small bug of wrong internal copying size.

### Verison 0.1.6.22

* Fixed bug of original source (ShuWang's SRCNN).

   - Use last layer (3) to Y channel at last construction.

* Changed ProcessSRCNN() method to get optional convolutional result.

### Verison 0.1.6.20

* Fixed memory leak after convolution55.

* Changed ProcessSRCNN() method to get convolutional gray.

### Version 0.1.4.17

* Bug fixed for color space conversion.

### Verison 0.1.5.18

* Supports variable filters for interpolation.

    1. Nearest

    1. Bilinear

    1. Bicubic

    1. Lanczos-3

    1. B-Spline

## License

* Follows original source GPLv2, but this project is LGPLv3.

## Requirements

* Your G++.

## How to build ?

* Make a symlink from `Makefile.{your platform}` in `makefiles` directory.

    - eg.) `ln -s makefiles/Makefile.macos Makefile`

* Then build with `make`.

* Testing applications may one of these,

	- `make -f makefiles/Makefile.test`

	- or

	- `make -f makefiles/Makefiles.testmac`

## Dependency

* Testing application by `make -f makefiles/Makefile.test` may requires FLTK and fl_imgtk libraries.

* FLTK should be installed by anyway, but recommend to my below FLTK-custom with fl_imgtk.

* [FLTK-custom](https://github.com/rageworx/fltk-custom) and [fl_imgtk](https://github.com/rageworx/fl_imgtk) for build test program for read and write image files.