import numpy as np
[docs]
def Default_Params():
"""
This function creates a set of parameters, stored in params, with default values
Returns:
- params: dictionary containing the different parameters
Author: Jerome Gilles
Institution: San Diego State University
Version: 1.0 (12/18/2024)
"""
params = {
# General parameters
"log": False, # if 1 the boundaries detection will be perform on the log of the spectrum
"N": 4, # Number of expected modes
"SamplingRate": 2*np.pi, # sampling frequency, if unknown set to 2pi for a normalized frequency line. i.e. [-pi,pi]
# Global trend parameters
"globtrend": "none", # options: none, powerlaw, polylaw, morpho, tophat, opening
"degree": 7, # degree of polynomial interpolation for polylaw
# Regularization parameters
"reg": "none", # options: none, gaussian, average, closing
"lengthfilter": 3, # length of filter
"sigmafilter": 1.5, # standard deviation for Gaussian regularization
# Boundary detection parameters
"detect": "scalespace", #detection methods: locmax, locmaxmin, locmaxminf, adaptivereg, adaptive, scalespace
"typeDetect": "otsu", #scale-space method: "otsu", "halfnormal", "empiricallaw", "mean", "kmeans"
"kn": 3, #scale kernel size for scalespace kernel
"t": 0.16, #initial scale
"InitBounds": np.array([5, 28, 54, 81]), #array of initial bounds to be adapted (adaptive and adaptivereg methods)
"Completion": False, # Request completion of the number of mode if less than N
# 1D transform
"wavname": "littlewood-paley", # mother wavelet: littlewood-paley, shannon, meyer, gabor1, gabor2
# Curvelet parameters
"option": 3, # specify which transform option: 1 = independent scales and angles, 2 = scales first then angles, 3 = angles first then scales
"curvdegree": 4, # degree of polynomial interpolation for polylaw for angles detection
"curvpreproc": "none", # options: none, powerlaw, polylaw, morpho, tophat, opening for angles detection
"curvreg": "none", # options: none, gaussian, average, closing for angles detection
"curvlengthfilter": 10, # length of filter for angles detection
"curvsigmafilter": 1.5, # standard deviation for Gaussian regularization for angles detection
"curvmethod": "scalespace", # detection methods: locmax, locmaxmin, scalespace for angles detection
"curvN": 6, # number of expected angular sector
# 2D Gaussian scale space extra parameters
"niter": 4, # number of iterations
"edge": 0, # size (in pixels) of the strip to ignore at the edge of the image
"includeCenter": 0, # if 1, the center of the image is included in the scale space maxima
# Voronoi and watershed partition parameters
"complex": 0, # if 1, the Voronoi partition is complex, otherwise it is real
"tau": 0.1 # half width of the transition area
}
return params