I have an RGB image -ndarray- and I want to count the occurrence of some colors like [255,0,0] or [0,0,255] in this image.
example of image data
np.ones((3, 3, 3)) * 255
array([[[255., 255., 255.],
[255., 255., 255.],
[255., 255., 255.]],
[[255., 255., 255.],
[255., 255., 255.],
[255., 255., 255.]],
[[255., 255., 255.],
[255., 255., 255.],
[255., 255., 255.]]])
So as result I want something like this
{
'[255,255,255]' : 9,
}
One solution could be the Counter function:
from collections import Counter
import numpy as np
# Generate some data
data = np.ones((10, 20, 3)) * 255
# Convert to tuple list
data_tuple = [ tuple(x) for x in data.reshape(-1,3)]
Counter(data_tuple)
Returns:
Counter({(255.0, 255.0, 255.0): 200})
Whereas its possible to use Counter or opencv histogram function to compute frequency of every single pixel , for specific pixels, its more efficient to use this:
import numpy as np
ar = np.ones([3,3,3]) *255
ar[1,1,:] = [0, 0, 200]
pixels = dict()
pixels['[255, 255, 255]'] = np.sum(np.all(ar == [255,255, 255], axis = 2))
pixels['[0, 0, 200]'] = np.sum(np.all(ar == [0, 0, 200], axis = 2))
result : {'[255, 255, 255]': 8, '[0, 0, 200]': 1}
Here's an approach using NumPy. Being values in the range 0-255, we could view the rows as tuples with three elements of type f8, and use np.unique to count the occurrences of the actual rows in the original ndarray. Using nakor's array:
a = np.ones((10, 20, 3)) * 255
We could then do:
vals, counts = np.unique(a.view('f8,f8,f8'), return_counts=True)
Where:
print(vals)
array([(255., 255., 255.)],
dtype=[('f0', '<f8'), ('f1', '<f8'), ('f2', '<f8')])
print(counts)
array([200])
ValueError: When changing to a larger dtype, its size must be a divisor of the total size in bytes of the last axis of the array.