Skip to content

normalize_points

normalize_points(v, center=None)

Make points fit into a side-length one cube

Translate and scale and arbitrary point set so that it's contained tightly into a 1 by 1 # (by 1) cube, centered at zero by default. Simple yet useful to test code without worrying about scale-dependencies.

Parameters:

Name Type Description Default
v (n,d) numpy double array

Matrix of point position coordinates

 required
center numpy double array (optional, None by default)

Where to center the mesh (if None, centered at zero)

 None

Returns:

Name Type Description
u numpy double array

Normalized point position coordinates

Examples:

# Generate a mesh
V, F = regular_square_mesh(3)
# By default, this mesh is centered at zero and has side length two
V = normalize_points(V,center=np.array([0.5,0.5]))
# Now it's centered at (0.5,0.5) and has side length one
Source code in src/gpytoolbox/normalize_points.py 
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
def normalize_points(v,center=None):
    """Make points fit into a side-length one cube

    Translate and scale and arbitrary point set so that it's contained tightly into a 1 by 1 # (by 1) cube, centered at zero by default. Simple yet useful to test code without worrying about scale-dependencies.

    Parameters
    ----------
    v : (n,d) numpy double array
        Matrix of point position coordinates
    center : numpy double array (optional, None by default)
        Where to center the mesh (if None, centered at zero)

    Returns
    -------
    u : numpy double array
        Normalized point position coordinates

    Examples
    --------
    ```python
    # Generate a mesh
    V, F = regular_square_mesh(3)
    # By default, this mesh is centered at zero and has side length two
    V = normalize_points(V,center=np.array([0.5,0.5]))
    # Now it's centered at (0.5,0.5) and has side length one
    ```
    """

    # Dimension:
    dim = v.shape[1]


    # First, move it to the first quadrant:
    for dd in range(dim):
        v[:,dd] = v[:,dd] - np.amin(v[:,dd])
    # Normalize for max length one    
    v = v/np.max(v)
    # Center at zero
    for dd in range(dim):
        v[:,dd] = v[:,dd] - 0.5*np.amax(v[:,dd])

    if (center is not None):
        v = v + np.tile(center,(v.shape[0],1))

    return v