def test_MTF(self):
"""
See that the MTF compares to Shi et al. and Star Lack et al.
Within four percent of shi and the star lack is far out now.
"""
s = fc.Spectrum()
s.load("W_spectrum_6") # Varian_truebeam')
kernel = fc.Detector(s, "CuGOS-336-micrometer")
kernel.get_plot_mtf_real(None)
shi_x = np.load(os.path.join("test_data", "shi_frequencies.npy"))
shi_y = np.load(os.path.join("test_data", "shi_MTF.npy"))
int_mtf = np.interp(shi_x, kernel.freq, kernel.mtf)
# print(np.max(np.abs(int_mtf - shi_y)))
# Within 10 percent
self.assertLess(np.max(np.abs(int_mtf - shi_y / shi_y[0])), 0.04)
# Star lack
star_x = np.load(os.path.join("test_data", "star_frequencies.npy"))
star_y = np.load(os.path.join("test_data", "star_MTF.npy"))
kernel2 = fc.Detector(s, "CWO-784-micrometer")
kernel2.get_plot_mtf_real(None)
int_mtf2 = np.interp(star_x, kernel2.freq, kernel2.mtf)
# print(np.max(np.abs(int_mtf2 - star_y/star_y[0])))
# Within 10 percent
self.assertLess(np.max(np.abs(int_mtf2 - star_y / star_y[0])), 0.20)
def test_MV_experimental_profile(self):
"""
Test that the results match an experimental
profile from Varian Truebeam within
0.1, This one for the MV image.
"""
spectra = ["Varian_truebeam"]
MV_detectors = ["CuGOS-784-micrometer"]
angles = np.linspace(np.pi / 2, np.pi * 2, 2)
phantom = fc.Catphan_404()
phantom.geomet.DSD = 1510
phantom.phan_map = [
"air",
"G4_POLYSTYRENE",
"G4_POLYVINYL_BUTYRAL",
"G4_POLYVINYL_BUTYRAL",
"CATPHAN_Delrin",
"G4_POLYVINYL_BUTYRAL",
"CATPHAN_Teflon_revised",
"air",
"CATPHAN_PMP",
"G4_POLYVINYL_BUTYRAL",
"CATPHAN_LDPE",
"G4_POLYVINYL_BUTYRAL",
"CATPHAN_Polystyrene",
"air",
"CATPHAN_Acrylic",
"air",
"CATPHAN_Teflon_revised",
"air",
"air",
"air",
"air",
]
s = fc.Spectrum()
s.load(spectra[0])
s.x[0] = 1
s.x[1] = 2
s.attenuate(0.0, fc.get_mu(z=13)) # 3.7
kernel = fc.Detector(s, MV_detectors[0])
kernel.add_focal_spot(1.4)
phantom.return_projs(kernel,
s,
angles,
mgy=0.0,
bowtie=True,
filter="FF_t0")
exp_profile = np.load(
os.path.join("test_data", "MV_experimental_profile.npy"))
self.assertLess(
np.mean(np.abs(np.roll(phantom.proj[0, 5], -3) - exp_profile)), 1)
def test_counts_comparison(self):
"""
Compares the counts in a profile for the 515
phantom between MC and Fastcat
3 Tests for the Al W and a kV spectrum
"""
s2 = fc.calculate_spectrum(120, 12, 3, 50, monitor=None)
s2.attenuate(0.1, fc.get_mu(z=13))
spectra = ["Al_spectrum_6", "W_spectrum_6", "kV"]
data_files = [
"MC_MV_6Al_counts",
"MC_MV_6xW_counts",
"MC_kV_120_counts",
]
angles = np.linspace(0, np.pi * 2, 2)
phantom = fc.Catphan_515()
s = fc.Spectrum()
for ii, spectrum in enumerate(spectra):
if spectrum == "kV":
s = s2
else:
s.load(spectrum)
kernel = fc.Detector(s, "CWO-784-micrometer")
# No noise special call of this function to return profile and not
# The projection which would be log(I/I_0)
counts = np.mean(
phantom.return_projs(kernel,
s,
angles,
return_intensity=True,
mgy=0.0)[0],
0,
)
MC_counts = np.load(
os.path.join("test_data", data_files[ii] + ".npy"))
self.assertLess(np.abs(np.max(counts - MC_counts)), 10)
def test_dose_al(self):
"""
Test if the dose matches MC estimates
Testing the C 2.5 and W 2.5 MV
"""
# Intialize
angles = np.linspace(0, 2 * np.pi, 2)
s_fc = fc.Spectrum()
phantom = fc.Catphan_515()
# Test the C_25
s_fc.load("C_spectrum_25")
kernel = fc.Detector(s_fc, "CuGOS-784-micrometer")
dose = phantom.return_projs(
kernel,
s_fc,
angles,
nphoton=2e7,
mgy=0,
return_dose=True,
verbose=0,
)
# print(dose[-1])
# C_25 test
self.assertAlmostEqual(dose[-1], 0.04099205, places=2)
# Test the W_25
s_fc.load("W_spectrum_6")
kernel = fc.Detector(s_fc, "CuGOS-784-micrometer")
dose = phantom.return_projs(
kernel,
s_fc,
angles,
nphoton=2e7,
mgy=0,
return_dose=True,
verbose=0,
)
# C_25 test
self.assertAlmostEqual(dose[-1], 0.1935003, places=2)
def test_no_residual(self):
"""
Making sure that the scan profile is flat without an object
And that the phantom reproduces the attenuation for water
"""
spectra = ["Varian_truebeam_phasespace"]
MV_detectors = ["CuGOS-784-micrometer"]
angles = np.linspace(np.pi / 2, np.pi * 2, 2)
phantom = fc.Catphan_404()
phantom.phan_map = ["1"] * 20
s = fc.Spectrum()
s.load(spectra[0])
s.x = np.array([290, 300, 310, 3000])
s.y = np.array([0, 1, 0, 0])
kernel = fc.Detector(s, MV_detectors[0])
kernel.deposition_interpolated = np.array([1, 0, 0])
phantom.return_projs(kernel,
s,
angles,
mgy=0.0,
scat_on=False,
convolve_on=False)
# Check that the profile is flat without object
self.assertLess(np.mean(phantom.proj[0]), 0.01)
# Check that the correct attenuation is returned
phantom.phan_map = ["water"] * 20
phantom.return_projs(kernel,
s,
angles,
mgy=0.0,
scat_on=False,
convolve_on=False)
self.assertAlmostEqual(np.max(phantom.proj[0]),
0.97 * 0.1186 * 20 * 10,
delta=0.1)