Balanced histogram thresholding: Difference between revisions

Revision as of 11:24, 6 October 2019

Revision as of 20:24, 30 November 2019

Line 38:

The following, is a possible implementation in the [[Python (programming language)|Python]] language:

<syntaxhighlight lang="python">

def bht(hist, min_count: int = 5) -> int:

"""Balanced histogram thresholding."""

n_bins = len(hist) # assumes 1D histogram

h_s = 0

while hist[h_s] < min_count:
h_s += 1 # ignore small counts at start

h_e = n_bins - 1

while hist[h_e] < min_count:
h_e -=1 1 # ignore small counts at end

# use mean intensity of histogram as center; alternatively: (h_s + h_e) / 2)

h_c = int(round(np.average(np.linspace(0, 2**8-1, n_bins), weights=hist)))

w_l = np.sum(hist[h_s:h_c]) # weight in the left part

w_r = np.sum(hist[h_c:h_e+1]) # weight in the right part

while h_s < h_e:

if w_l > w_r: # left part became heavier

w_l -= hist[h_s]

Line 64 ⟶ 67:

w_l += hist[h_c]

w_r -= hist[h_c]

h_c = new_c