codeare
1.2

This is the developer's guide and API reference to codeare, the common data exchange and reconstruction platform. If you are looking for the general documentation, tutorials etc, please refer to http://codeare.org
codeare is most readily devided in three major realms:
In a nutshell, all three parts together allow one to have parts of a live or retrospective reconstruction or runtime feedback run outside your vendor's reconstrcution chain. It could be, however also used to have an external device, let's say a camera, an HID device and vice versa, to interact with a running sequence. It could on the other hand be used to engange with programming languages / paradigms that may not be possible with your vendor's setup; think of using python, MATLAB, maple, Mathematica, ... LIVE!
The matrix template is essentially the template for an NDarray that comes with an extensive and extensible toolbox to allow one to conentrate entirely on the implementation of the algorithm in question. Most of the commands that act on matrices are very much similiar if not identlical to the way one deals with matrices in python, MATLAB and octave etc. (f.e. A = fft(A)
);
Implementationwise the major difference to the mentioned programming environments and languages, lies in the use of references rather than copies of matrices for memory efficiency and the fact that the core implementation is inherently bound to the C++ language specifications. So that [u,s,d] = svd (a)
is not option. The SVD case is handled here as
svd(const Matrix<T>& IN, Matrix<S>& s, Matrix<T>& U, Matrix<T>& V, const char& jobz = 'N').