. 2020 Jul 7;7:297. doi: 10.3389/fmed.2020.00297

Algorithm 1.

Fuzzy concept creation

	input : grayscale image I
	output: fuzzy set (concept)
1	For the fuzzy concepts Bright, Dark, and Median, specify the intervals of [Bright_begin, Bright_end], [Dark_begin, Dark_end], and [Median_begin, Median_end], respectively.
2	Let $B r i g h t_{b e g i n} = (N_{f} - 1) [\frac{L - 1}{N_{f}}]$ , $D a r k_{e n d} = [\frac{L - 1}{N_{f}}]$ , Median_begin = Dark_end−left_overlap, and Median_end = Bright_begin+right_overlap /* N_f is the number of fuzzy concepts, and left_overlap and right_overlap determine the overlapping range of the fuzzy concepts (the overlap range was set to 0 in this paper) */.
3	Set Dark_begin to be the first g_k from 0 to Dark_end.
4	Set Bright_end to be the last g_k from Bright_begin to L − 1.
5	In interval [Dark_begin, Dark_end], find a pixel g_k with the maximum value of p(g_k).
6	For the fuzzy concept Dark, create its membership function f_Dark as follows: m_Dark ← g_k, α_Dark ← m_Dark − Dark_begin, β_Dark ← Dark_end − m_Dark.
7	In interval [Median_begin, Median_end], find a pixel g_k with the maximum value of p(g_k).
8	For the fuzzy concept Median, create its membership function f_Median as follows: m_Median ← g_k, α_Median ← m_Median − Median_begin, β_Median ← Median_end − m_Median.
9	In interval [Bright, begin − Bright_end], find a pixel g_k with the maximum value of p(g_k).
10	For the fuzzy concept Bright, create its membership function f_Bright as follows: m_Bright ← g_k, α_Bright ← m_Bright − Bright_begin, β_Bright ← Bright_end − m_Bright.