| 156 |
jab |
1 |
/** \mainpage GEL: GEometric and Linear algebra tools.
|
|
|
2 |
|
|
|
3 |
GEL is a framwork for computer graphics and 3D vision.
|
|
|
4 |
|
|
|
5 |
There are a good many tools for computational geometry processing: A
|
| 211 |
jab |
6 |
voxel grid and a simple triangle mesh data structure. There is also
|
|
|
7 |
a halfedge based polygonal mesh data structure called HMesh. There is a very
|
| 156 |
jab |
8 |
useful data structure known as a k-D tree and many other things.
|
|
|
9 |
|
|
|
10 |
Also found are two packages for linear algebra: CGLA is strictly for small
|
|
|
11 |
vectors and matrices. LinAlg is a Lapack wrapper for slightly larger
|
|
|
12 |
problems. At this point, it is fairly simple.
|
|
|
13 |
|
|
|
14 |
Finally there are some utilities in Util: Getting command line arguments,
|
|
|
15 |
hash table classes, a 2D grid class, a resource manager and other
|
|
|
16 |
miscellany.
|
|
|
17 |
*/
|
|
|
18 |
|
|
|
19 |
/** \namespace CGLA
|
|
|
20 |
\brief Computer Graphics Linear Algebra.
|
|
|
21 |
|
|
|
22 |
CGLA is a set of numerical C++ vector and matrix classes and class
|
|
|
23 |
templates designed with computer graphics in mind. CGLA stands for
|
|
|
24 |
``Computer Graphics Linear Algebra''.
|
|
|
25 |
|
|
|
26 |
Let us get right down to the obvious question: Why create another
|
|
|
27 |
linear algebra package?
|
|
|
28 |
Well, CGLA evolved from a few matrix and vector classes because I
|
|
|
29 |
didn't have anything better. Also, I created CGLA to experiment with
|
|
|
30 |
some template programming techniques. This led to the most important
|
|
|
31 |
feature of CGLA, namely the fact that all vector types are derived
|
|
|
32 |
from the same template.
|
|
|
33 |
|
|
|
34 |
This makes it easy to ensure identical semantics: Since all vectors
|
|
|
35 |
have inherited, say, the * operator from a common ancestor, it works
|
|
|
36 |
the same for all of them.
|
|
|
37 |
|
|
|
38 |
It is important to note that CGLA was designed for Computer Graphics
|
|
|
39 |
(not numerical computations) and this had a number of
|
|
|
40 |
implications. Since, in computer graphics we mainly need small vectors
|
|
|
41 |
of dimension 2,3, or 4 CGLA was designed for vectors of low
|
|
|
42 |
dimensionality. Moreover, the amount of memory allocated for a vector
|
|
|
43 |
is decided by its type at compile time. CGLA does not use dynamic
|
|
|
44 |
memory. CGLA also does not use virtual functions, and most functions
|
|
|
45 |
are inline. These features all help making CGLA relatively fast.
|
|
|
46 |
|
|
|
47 |
Of course, other libraries of vector templates for computer graphics
|
|
|
48 |
exist, but to my knowledge none where the fundamental templates are
|
|
|
49 |
parametrized w.r.t. dimension as well as type. In other words, we have
|
|
|
50 |
a template (ArithVec) that gets both type
|
|
|
51 |
(e.g. float) and dimension
|
|
|
52 |
(e.g. 3) as arguments. the intended use of this template is as
|
|
|
53 |
ancestor of concrete types such as Vec3f - a 3D floating
|
|
|
54 |
point type.
|
|
|
55 |
|
|
|
56 |
The use of just one template as basis is very important, I believe,
|
|
|
57 |
since it makes it extremely simple to add new types of
|
|
|
58 |
vectors. Another very generic template is ArithMat which is a
|
|
|
59 |
template for matrix classes. (and not necessarily NxN matrices).
|
|
|
60 |
|
|
|
61 |
From a users perspective CGLA contains a number of vector and matrix
|
|
|
62 |
classes, a quaternion and some utility classes. In summary, the most
|
|
|
63 |
important features are
|
|
|
64 |
|
|
|
65 |
- A number of 2, 3 and 4 d vector classes.
|
|
|
66 |
- A number of Matrix classes.
|
|
|
67 |
- A Quaternion class.
|
|
|
68 |
- Some test programs.
|
|
|
69 |
- Works well with OpenGL.
|
|
|
70 |
|
|
|
71 |
|
|
|
72 |
There is a document on CGLA in the GEL documentation. The introduction
|
|
|
73 |
above was taken from that text.
|
|
|
74 |
*/
|
|
|
75 |
|
|
|
76 |
/** \namespace Graphics
|
|
|
77 |
\brief The namespace for things related to (real time) rendering
|
|
|
78 |
|
|
|
79 |
Right now this namespace contains mostly trackball and viewing related
|
|
|
80 |
things.
|
|
|
81 |
*/
|
|
|
82 |
|
|
|
83 |
/** \namespace Geometry
|
|
|
84 |
\brief A namespace for utilities related to geometry.
|
|
|
85 |
|
|
|
86 |
This namespace contains a wide range of stuff: Spatial datastructures
|
|
|
87 |
voxel grids and related classes as well as a simple triangle mesh class.
|
|
|
88 |
*/
|
|
|
89 |
|
|
|
90 |
/** \namespace HMesh
|
|
|
91 |
\brief The HMesh namespace contains the Manifold class which is a halfedge
|
|
|
92 |
based mesh.
|
|
|
93 |
|
|
|
94 |
Apart from manifold there are also face and vertex circulators in this
|
|
|
95 |
namespace. More advanced things are relegated to the HMeshUtil namespace.
|
|
|
96 |
|
|
|
97 |
Some applications are also found here. For instance an isosurface
|
|
|
98 |
polygonizer and Garland Heckbert simplification has been implemented
|
|
|
99 |
on top of HMesh.
|
|
|
100 |
*/
|
|
|
101 |
|
|
|
102 |
/*!
|
|
|
103 |
\namespace LinAlg
|
|
|
104 |
\brief The Linear Algebra Implementation/Module/Package.
|
|
|
105 |
|
|
|
106 |
This is the linear algebra package deal with vector and matrix types plus
|
|
|
107 |
the basic operation on them. The Vector and Matrix Types are templated such
|
|
|
108 |
that the precision of a given 'system' can be changen. Though the precision
|
|
|
109 |
or type in mind is the double. Special types for 2 and 3 vectors
|
|
|
110 |
(CVec2 and CVec3) are included for improved performance. The more advanced
|
|
|
111 |
linear algebra functions are supplied by lincage to the LaPack package.
|
|
|
112 |
|
|
|
113 |
\test The functionality in this pacage has if nothing else is mentioned
|
|
|
114 |
been tested by making small functions utilizing the functions. The
|
|
|
115 |
reason being that the functions here are rather esay to validate due to
|
|
|
116 |
their functional simplicity.
|
|
|
117 |
|
|
|
118 |
A list of improvements in mind is:
|
|
|
119 |
- Lapack interface
|
|
|
120 |
- Vector and Matrix product and so on, with interface to _BLAS_
|
|
|
121 |
- iterator thing
|
|
|
122 |
- Memory leak check ? inline ?
|
|
|
123 |
- Element functor
|
|
|
124 |
- Identity matrixs
|
|
|
125 |
- Compile and utilize the newer versions of the Lapack routines.
|
|
|
126 |
- Have a nameing convention section in the documentaion
|
|
|
127 |
- Integrate the design into the documentation
|
|
|
128 |
- Have a get data in the matrix and vector instead of &A[0].
|
|
|
129 |
- Have 3x3, 3x4 4x1 and 4x4 specialities.
|
|
|
130 |
- Look at casting e.g. vec2i=vec2l.
|
|
|
131 |
- Error handling in From/To matrix
|
|
|
132 |
- Has error handling in Lapack been made ?
|
|
|
133 |
|
|
|
134 |
\author Henrik Aanęs
|
|
|
135 |
\version Aug 2001
|
|
|
136 |
*/
|
|
|
137 |
|
|
|
138 |
/** \namespace Util
|
|
|
139 |
\brief This namespace is for general utilities that do not fit elsewhere.
|
|
|
140 |
|
|
|
141 |
*/
|
|
|
142 |
|