Research Article

Blind Source Parameters for Performance Evaluation of Despeckling Filters

Table 3

Input parameters of despeckling techniques.

Method(s)Parameters

DsFlsminsi and DsFlsmv [5, 22]Window size = 5 × 5; iterations = 2, also with 3 × 3, 7 × 7, and 9 × 9
FBL [20]Width of spatial Gaussian = 10; width of range Gaussian = 20
Fuzzy filter [21, 24]Window size = 3 × 3; padding, also with 5 × 5, 7 × 7, and 9 × 9
FIF/HFIF and FBF/HFBF [21, 24] = 500, with order = 2, also with = 100 and 1000
HDWT and HDTDWT [3]Window size = 3 × 3; level = 2
DsFhomog [5, 22]Window = 3 × 3
OBNLM [4, 14]Search area = 23 × 23; block size = 15 × 15; smoothing value = 0.4
Lee, Kaun, and Frost [1, 37]Window size = 5 × 5
ProbShrink [30]Window size = 3 × 3; level = 2; sym8
AFTV [16]Iterations = 3; λ = 0; time step Δt = 0.2
SURELET [29]Downsampling = 4; overlap factor = 3; redundancy = 3
RNLA [11, 12]Support = 1; degree = 3
NSS [28]Level = 3; sym8
MBR [9]Band M = 3; = 5000
ATV [17]Iterations = 2; λ = 1; time step Δt = 0.2
DsFad [1]Diffusion constant = 30; rate of diffusion = 0.25; iterations = 20
DsFsrad [1, 5]Iterations = 30; time step Δt = 0.02; = 1
CED [1]Iterations = 20; time step Δt = 0.02; diffusion constant = 20
DPAD [4, 34]Iterations = 30; time step Δt = 0.02; Cu noise estimation
DsFgf4d [5, 22, 35]Window = 3 × 3; iterations = 2
DsFmedian [5, 22]Window = 5 × 5; iterations = 3
PPB [4, 15]Iterations = 4; α = 0.8; = 2; search area = 23 × 23; patch size = 7 × 7
DsFhmedian [5, 22]Window = 5 × 5; iterations = 2
Hybrid fuzzy [21]Fuzzy and Wiener window size = 3 × 3
ROF [18]Time step Δt = 0.25; iterations = 5
PSBE [10] Spatial sigma = 0.01; window size = 21 × 21; samples = 100
MPT [27]Scale number = 2; = 12
BayesShrink [25] wtype = db4; level = 2
GLM [13]Window size = 3 × 3; level = 2; = 3
DsFwiener [5, 22]Window size = 5 × 5, 3 × 3, and 7 × 7

Spatial sigma refers to spatial variance of the initial probability distribution; wtype refers to wavelet type.