Mark pixels of a blank image-CodePudding

I Want to mark the pixels,

Mark=[2, 455, 6, 556, 12, 654, 22, 23, 4,86,.....] in such a way that it will not mark the 1st 2 pixels and then mark next 455 pixels by a color, again for next 6 pixels it will not mark and again mark the next 556 pixels by the same color and so on. The size of the image is 500x500x3. How do I calculate these steps?

Img=np.zeros((500,500,3),dtype=np.uint8)

CodePudding user response：

Your algorithm is actually in your question. By 500x500x3 I guess that you mean your image is 500 (width) on 500 (height) with 3 color channel?

It could be implemented as follows, without any optimizations:

color = (128, 50, 30)
x, y = 0, 0
for (skip, count) in zip([a for i, a in enumerate(Mark) if i % 2 == 0], [a for i, a in enumerate(Mark) if i % 2 != 0]):
    x  = skip
    y  = x // 500  # keep track of the lines, when x > 500,
                       # it means we are on a new line
    x %= 500  # keep the x in bounds

    # colorize `count` pixels in the image
    for i in range(0, count):
        
        Img[x, y, 0] = color[0]
        Img[x, y, 1] = color[1]
        Img[x, y, 2] = color[2]
        x  = 1
        y  = x // 500
        x %= 500  # keep the x in bounds

The zip([a for i, a in enumerate(Mark) if i % 2 == 0], [a for i, a in enumerate(Mark) if i % 2 != 0]) is a just a way to group the pairs (skip, pixel to colorize). It could definitely be improved though, I'm no Python expert.

CodePudding user response：

The easiest way is probably to convert the image to a Numpy array:

import numpy as np
na = np.array(Img)

And then use Numpy ravel() to give you a flattened (1-D) view of the array

flat = np.ravel(na)

You can now see the shape of your flat view:

print(flat.shape)

Then you can do your colouring by iterating over your array of offsets from your question. Then the good news is that, because ravel() gives you a view into your original data all the changes you make to the view will be reflected in your original data.

So, to get back to a PIL Image, all you need is:

RecolouredImg = Image.fromarray(na)

Try it out by just colouring the first ten pixels before worrying about your long list.

If you like working with Python lists (I don't), you can achieve a similar effect by using PIL getdata() to get a flattened list of pixels and then process the list against your requirements and putdata() to put them all back. The effect will be the same, just choose a method that fits how your brain works.