Making figure transparent with colored background

Question

I have a bit of a situation. What I need is a plot with a black background with several white circles drawn on top of that black background.

I managed to do this usi

Answer 1

Circles masked with colorbar

Colormaps can have an alpha channel, so if your data is on a mesh with high vs low values showing circle vs not-circle, one set of those values can be transparent.

This only works for me when saving the figure programmatically, with the transparent keyword; not from the Python image window.

Starting from one of the matplotlib gallery examples (in a gimp-alike, I can cut & paste segments and the transparency is right):

# plot transparent circles with a black background
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.colors import LinearSegmentedColormap
from matplotlib.cm import Greys


dark_low = ((0., 1., 1.),
         (.3, 1., 0.),
         (1., 0., 0.))
         
cdict = {'red':  dark_low,

     'green': dark_low,

     'blue': dark_low}

cdict3 = {'red':  dark_low,

     'green': dark_low,

     'blue': dark_low,
               
     'alpha': ((0.0, 0.0, 0.0),
               (0.3, 0.0, 1.0),
               (1.0, 1.0, 1.0))
    }

greys = LinearSegmentedColormap('Greys', cdict)
plt.register_cmap(cmap=greys)
    
dropout_high = LinearSegmentedColormap('Dropout', cdict3)
plt.register_cmap(cmap = dropout_high)

# Make some illustrative fake data:

x = np.arange(0, np.pi, 0.1)
y = np.arange(0, 2*np.pi, 0.1)
X, Y = np.meshgrid(x,y)
Z = np.cos(X) * np.sin(Y) * 10

# Make the figure:

plt.figure()
plt.subplot(1,3,1)
plt.imshow(Z, cmap=Greys)
plt.title('Smooth\ncolorbar')
plt.colorbar()
plt.subplot(1,3,2)
plt.imshow(Z, cmap=greys)
plt.title('Linear\ncolorbar')
plt.colorbar()
plt.subplot(1,3,3)
plt.imshow(Z, cmap = dropout_high)
plt.title('Alpha crops\n colorbar')
plt.colorbar()
plt.savefig('dropout_cmap', transparent=True)

And as a layer over another image. Interesting, the colorbar with alpha channel doesn't have transparency. That seems like a bug.

Answer 2

This might not be the answer you are looking for but it gives the picture you wanted! I think you want to fill the areas outside the circle!(s) with black and leave the background transparent, rather than the other way around. It's trivial to calculate the boundaries of a single circle and use fill_between. Doing it for multiple circles might be trickier!

import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure()
ax = fig.add_subplot(1, 1, 1, aspect = "equal", )

# A Circle
xy=(1,1); r=3

# more points is smoother
xdata=np.linspace(-5,5,1001)
ydata=np.linspace(-5,5,1001)

# circle edges (top and bottom)
c1=np.sqrt((xy[0]**2-xdata**2)+r**2)+xy[1]
c2=-np.sqrt((xy[0]**2-xdata**2)+r**2)+xy[1]

c1=np.where(np.isnan(c1),xy[0],c1)
c2=np.where(np.isnan(c2),xy[0],c2)

ax.fill_between(xdata,5,c1,color='black')
ax.fill_between(xdata,-5,c2,color='black')

plt.xlim(-5, 5)
plt.ylim(-5, 5)

fig.savefig("test.png", dpi = 300, transparent=True)

Transparent circle with center 1,1 and radius 3

Answer 3

Edit 3:

It has been clarified that the underlying question is:

how to put a 'black & transparent' mask in front of a matplotlib image produced by imshow ? The mask shall result from a matplotlib previously drawn black & white figure.

The following code demonstrate this feature by accessing and mixing the figure rgba bitmaps:

import numpy as np
import matplotlib.pyplot as plt

import matplotlib.cm as cm
import matplotlib.mlab as mlab


def get_rgba_bitmap(fig):
    fig.canvas.draw()
    tab = fig.canvas.copy_from_bbox(fig.bbox).to_string_argb()
    ncols, nrows = fig.canvas.get_width_height()
    return np.fromstring(tab, dtype=np.uint8).reshape(nrows, ncols, 4)

def black_white_to_black_transpa(rgba):
    rgba[:, :, 3] = 255 - rgba[:, :, 0]
    rgba[:, :, 0:3] = 0

def over(rgba1, rgba2):
    if rgba1.shape != rgba2.shape:
        raise ValueError("rgba1 and rgba2 shall have same size")
    alpha = np.expand_dims(rgba1[:, :, 3] / 255., 3)
    rgba =  np.array(rgba1 * alpha + rgba2 * (1.-alpha), dtype = np.uint8)
    return rgba[:, :, 0:3]


# fig 1)
fig1 = plt.figure(facecolor = "white")
fig1.set_dpi(300)
ax1 = fig1.add_subplot(1, 1, 1, aspect = "equal", axisbg = "black")
ax1.add_artist(plt.Circle((0., 0., .5), color =   "white")) 
ax1.set_xlim(-5, 5)
ax1.set_ylim(-5, 5)
bitmap_rgba1 = get_rgba_bitmap(fig1)
black_white_to_black_transpa(bitmap_rgba1)

# fig 2
fig2 = plt.figure(facecolor = "white")
fig2.set_dpi(300)
delta = 0.025
ax2 = fig2.add_subplot(1, 1, 1, aspect = "equal", axisbg = "black")
ax2.set_xlim(-5, 5)
ax2.set_ylim(-5, 5)
x = y = np.arange(-3.0, 3.0, delta)
X, Y = np.meshgrid(x, y)
Z1 = mlab.bivariate_normal(X, Y, 1.0, 1.0, 0.0, 0.0)
Z2 = mlab.bivariate_normal(X, Y, 1.5, 0.5, 1, 1)
Z = Z2-Z1  # difference of Gaussians
im = ax2.imshow(Z, interpolation='bilinear', cmap=cm.jet,
                origin='lower', extent=[-5, 5, -5, 5],
                vmax=abs(Z).max(), vmin=-abs(Z).max())
bitmap_rgba2 = get_rgba_bitmap(fig2)    

# now saving the composed figure 
fig = plt.figure()
fig.patch.set_alpha(0.0) 
ax = fig.add_axes([0., 0., 1., 1.])
ax.patch.set_alpha(0.0) 
ax.imshow(over(bitmap_rgba1, bitmap_rgba2))
plt.axis('off')
fig.savefig("test_transpa.png", dpi=300)

plt.show()

Giving:

I tested with your initial photonic test case and the pic quality seems OK

Now if you want the figure background transparent too:

• Set fig1 background to 'white' i.e. fig1 = plt.figure(facecolor='white'), as white will become transparent when passed to black_white_to_black_transpa
• Set fig2 background to transparent fig2.patch.set_alpha(0.0) as it will be stored with no modification into bitmap_rgba2
• Finally, take care of the alpha channel when mixing bitmap_rgba1 and bitmap_rgba2 inside over function (see below a possible modification)

def over(rgba1, rgba2):
    if rgba1.shape != rgba2.shape:
        raise ValueError("rgba1 and rgba2 shall have same size")
    alpha1 = np.expand_dims(rgba1[:, :, 3] / 255., axis=3)
    alpha2 = np.expand_dims(rgba2[:, :, 3] / 255., axis=3)
    alpha = 1. - (1.-alpha1) * (1.-alpha2)
    C1 = rgba1[:, :, 0:3]
    C2 = rgba2[:, :, 0:3]
    C = (alpha1 * C1 + (1-alpha1) * alpha2 * C2) / alpha
    rgba =  np.empty_like(rgba1, dtype = np.uint8)
    rgba[:, :, 0:3] = C
    rgba[:, :, 3] = 255 * alpha[:, :, 0]
    return rgba

last (?) edit: It seems there is an inconsistence between the array returned byto_string_argb and the one expected by imshow (order of the rgb channels). A possible solution is to change ax.imshow(over(bitmap_rgba1, bitmap_rgba2)) to:

over_tab = over(bitmap_rgba1, bitmap_rgba2)
over_tab[:, :, 0:3] = over_tab[:, :, ::-1][:, :, 1:4]
ax.imshow(over_tab)