Table of Contents
Chapter Objectives
By reading this chapter, you will get to know
the basic ML features (see Section 1.1, “Overview”),
the concepts used in the ML (see Section 1.2, “Principles”),
a survey of ML classes:
the classes used for ML development (see Section 1.3.1, “Classes for Module Development”),
the administrative classes (see Section 1.3.2, “Administrative Classes”),
the image classes (see Section 1.3.3, “Image Classes”),
the helper classes (see Section 1.3.4, “Helper Classes”),
the APIs and classes for interaces/voxel type extension (see Section 1.3.5, “APIs and Classes for Interfaces and Voxel Type Extensions”),
the component groups (see Section 1.3.6, “Component Groups”),
and a short overview of modules already implemented or available in the ML Module database (see Section 1.3.7, “The ML Module Database”).
The MeVis Image Processing Library (called ML in the following) represents a general approach to image processing. It is based on the following principal ideas:
Image processing algorithms are represented by modules (sometimes also called operators or nodes).
Modules are mainly arranged in a directed graph that represents the flow of image data.
Modules implement a unified image processing interface.
Modules are self-descriptive by exporting their parameters as fields (field interface).
Image data is processed in fractions, i.e., page by page (paging).
Pages can be processed in parallel if supported by a module (multithreading).
Image processing is performed on a request-oriented basis (pull model, processing on demand).
Images can have up to six dimensions.
Image pixels (called voxels in this document) can be single scalars or structures (e.g vectors, complex values or matrices).
Platform independence, pure C++ code running on Windows, Linux and Mac OS X.
A C interface of the ML is available for applications that do not use C++.
© 2024 MeVis Medical Solutions AG