1

I'm using Python to count the frequency of pixel colors in an image. The Python Imaging Library can convert an image to a list of RGB values, and from there I can easily count duplicates, ending up with a dictionary of pixel values (as strings) and frequencies, like so:

{
    "255-255-255": 450,
    "255-254-254": 345,
    "249-250-255": 184,
    "124-130-200": 3,
} [etc etc]

(Essentially it's a histogram.)

For large images, I'm then quantizing the colors to multiples of N, so then I might have:

[
    ("255-255-255", 450),
    ("255-255-255", 345),
    ("250-250-255", 184),
    ("125-130-200", 3),
] [etc etc]

This leaves a lot of duplicate "keys" (stored as tuples since we have duplicates). I now need to condense, adding the values of all duplicates. So far I have:

c = 0
while c < len(vals) - 1:
    if vals[c][0] == vals[c+1][0]:
        vals[c][1] += vals[c+1][1]
        vals.pop(c+1)
    else:
        c += 1
return vals

It works fine, but there must be a way with list comprehensions? Or some other more efficient manner? I realize PIL may be able to do this, but I'd like to do by hand while learning how images work. Thanks!

4

1 回答 1

1

尝试这个：

l = [("255-255-255", 450),
    ("255-255-255", 345),
    ("250-250-255", 184),
    ("125-130-200", 3)]

from collections import defaultdict
D = defaultdict(int)

for k,v in l:
    D[k] += v

print D # display the dict.

>>> 
defaultdict(<type 'int'>, {'125-130-200': 3, '250-250-255': 184, '255-255-255': 795})
于 2013-05-27T15:59:41.137 回答