DIPimage User Manual » Introduction

The DIPimage toolbox

MATLAB is a software package designed for (among other things) data processing. It contains a huge amount of numerical algorithms, and very good data-visualization abilities. This makes it adequate for image processing. However, MATLAB’s virtues do not end there. It is also an ideal tool for rapid prototyping, since it handles a compact but simple notation and it is very easy to add functions to it. The drawback is that MATLAB, since it is an interpreted language, is slow for some constructs like loops; it also is not very efficient with memory (for example, all MATLAB data uses 8-byte floats). This makes it, by itself, a bit less useful beyond the prototyping stage.

DIPimage is a MATLAB toolbox for quantitative image analysis, and is based on the C++ library DIPlib. It is meant as a tool for research and development, as well as teaching image processing at various levels. This toolbox is made with user-friendliness, ease of implementation of new features, and compactness of notation in mind. The toolbox contains many efficient algorithms, though we always prioritize precision over speed. This means that, even though this toolbox contains many very efficient algorithms, there might be other (non-MATLAB) alternatives to think of if speed is your number one priority.

The DIPlib library

DIPlib is a quantitative image analysis library written in C++. It contains a large number of functions for processing and analyzing multi-dimensional image data. The library provides functions for performing transforms, filter operations, object generation, and statistical analysis of images. It is also very efficient (with both memory and time).

DIPimage, the MATLAB interface to DIPlib is more than a simple “glue” layer, often changing the syntax of the DIPlib function calls to be more natural within the MATLAB environment. Nonetheless, the help text for functions will list the DIPlib function that is called. The on-line DIPlib reference for the function can then be read to learn more details about the algorithm and the meaning of the parameters.

More information on DIPlib can be found at diplib.org and its documentation.

Image Analysis

This manual is meant as an introduction and reference to the DIPimage toolbox, not as a tutorial on image analysis. Although Getting Started shows some image analysis basics, it is not complete. We refer to “The Fundamentals of Image Processing”.

Documentation conventions

The following conventions are used throughout this manual:

• Example code: in typewriter font

• File names: in typewriter font

• Function names/syntax: in typewriter font

• Keys: like this

• Mathematical expressions: in italic

• Menu names, menu items, and controls: “inside quotes”

Acknowledgments

DIPlib was written mainly by Michael van Ginkel and Geert van Kempen, at the (then called) Pattern Recognition Group of Delft University of Technology. Cris Luengo rewrote the infrastructure in C++ for the 3.0 release.

The original DIPimage toolbox was written mainly by Cris Luengo, Lucas van Vliet, Bernd Rieger and Michael van Ginkel, also at Delft University of Technology. Tuan Pham, Kees van Wijk, Judith Dijk, Geert van Kempen and Peter Bakker contributed functionality. The 3.0 release was a rewrite of most components, to use the new version of the DIPlib library, by Cris Luengo.