The CImg Library - C++ Template Image Processing Toolkit

What is CImg?

The CImg Library is a small and open-source C++ toolkit for image processing, designed with these properties in mind :

	CImg defines classes and methods to manage images in your own C++ code. You can use CImg to load/save various file formats, access pixel values, display/transform/filter images, draw primitives (text, faces, curves, 3d objects, ...), compute statistics, manage user interactions on images, and so on...
	CImg defines a single image class able to represent datasets having up to 4-dimensions (from 1d scalar signals to 3d hyperspectral volumetric images), with template pixel types (`bool,char,int,float,...`). It also handles image collections and sequences.
	CImg is self-contained, thread-safe and highly portable. It fully works on different operating systems (`Unix,Windows,MacOS X,BSD,...`) and is compatible with various C++ compilers* (`Visual C++,g++,clang++,icc,...`).
	CImg is lightweight. It is made of a single header file `CImg.h` that must be included in your C++ source. It defines only four different classes, encapsulated in the namespace `cimg_library`. It can be compiled using a minimal set of standard C++ and system libraries only. No need for exotic or complex dependencies.
	Although not mandatory, CImg can use functionalities of external tools/libraries such as Board, FFMPEG, FFTW3. GraphicsMagick, ImageMagick, Lapack, libcurl, libjpeg, libpng, libtiff, Magick++, OpenEXR OpenCV OpenMP or XMedCon. Moreover, a simple plug-in mechanism allows any user to directly enhance the library capabilities according to his needs.
	CImg is a free, open-source library distributed under the CeCILL-C (close to the GNU LGPL) or CeCILL (compatible with the GNU GPL) licenses. It can be used in commercial applications.

CImg stands for Cool Image : It is easy to use, efficient and is intended to be a very pleasant toolbox to design image processing algorithms in C++. Due to its generic conception, it can cover a wide range of image processing applications.

Authors

David Tschumperlé (project leader)

(Check out his blog)

with the help of many contributors around the world :

Maksim Aizenshtein, Alberto Albiol, Antonio Albiol, Kiko Albiol, Simon Barthelmé, Neil Brown, Haz-Edine Assemlal, Vincent Barra, Wolf Blecher, Romain Blei, Yohan Bentolila, Jerome Boulanger, Pierre Buyssens, Sebastien Coudert, Frederic Devernay, Olivier D'Hondt, François-Xavier Dupé, Gerd von Egidy Eric Fausett, Jean-Marie Favreau, Sebastien Fourey, Alexandre Fournier, Vincent Garcia, David Grimbichler, Jinwei Gu, Jean-Daniel Guyot, Cédric Hammiche, Matt Hanson, Sebastien Hanel, Michael Holroyd, Christoph Hormann, Hon-Kwok Fung, Werner Jainek, Vo Duc Khanh, Daniel Kondermann, Pierre Kornprobst, Jan W. Krieger, Orges Leka, Francois Lauze, Xie Long, Thomas Martin, Cesar Martinez, Jean Martinot, Arnold Meijster (Center for High Performance Computing and Visualization, University of Groningen/The Netherlands) Nikita Melnichenko, Baptiste Mougel, Julien Morat, Jovana Milutinovich, Guillaume Nee, Adam Newgas, Francisco Oliveira, Andrea Onofri, Renaud Peteri, Martin Petricek, Paolo Prete, Adrien Reboisson, Klaus Schneider, Jakob Schluttig, Veronique Souchaud, Konstantin Spirin, Rainer Steffens, David G. Starkweather, Elle Stone, Grzegorz Szwoch, Thierry Thomas, Yu-En-Yun, Ingo Weyrich, Phillip Wood, Bug Zhao, Haibo Zheng.

Licenses

The CImg Library is an open-source product distributed under two distinct licenses : the library core itself is dual-licensed and can be governed either by the CeCILL-C License (LGPL-like), or the CeCILL License (GPL-compatible). Most of the other package files are distributed under the CeCILL License. Both are open-source licenses, the CeCILL-C being less restrictive than the CeCILL.

The CImg Library source code has been registered to the APP (French Agency for the Protection of Programs) by the INRIA, under registration number IDDN.FR.001.040004.000.S.P.2004.000.21000.

Learn more

The Tutorial section shows the basic use of CImg classes and functions with a small first code.
The Screenshots section illustrates some of the different source code examples provided in the CImg package.
You can look at the quite complete CImg Library presentation slides (.pdf format) which gives more insights on the different library concepts.
Online web statistics on the CImg website activity are available here.

Quick history

The development of the CImg Library began at the end of 1999, when I started my PhD thesis in the Lab at the Sophia Antipolis. It was designed to help me and my colleagues developing various image processing algorithms, for datasets as simple as 2D scalar images, or as complex as 3D volumes of diffusion tensors. I also used it for courses on image processing I teached at the university. As a result, the CImg Library has been always intended to be compact, easy to install and to use, multi-platform and generic. It provides a lot of basic functions that everyone would like to find in a good C++ image processing framework.

I am now a permanent researcher of the institution, working in the group at the lab in Caen/France. I am still using, maintaining and updating the CImg Library, and will probably do it for the next couple of years.