def Create_Ideal_DCE(input_folder, output_filepath = '', input_aif=''):
input_DCEs = []
input_niis = glob.glob(os.path.join(input_folder, '*nrrd'))
for nii in input_niis:
if 'ktrans' in nii or 've' in nii:
continue
else:
input_DCEs += [nii]
for input_4d_nifti in input_DCEs:
print 'Regenerating... ', input_4d_nifti
# if output_filepath == '':
output_filepath = str.split(input_4d_nifti, '.')[0]
input_ktrans = output_filepath + '_ktrans.nii.gz'
input_ve = output_filepath + '_ve.nii.gz'
input_4d_nifti = Convert_NRRD_to_Nifti(input_4d_nifti, input_ktrans)
input_numpy_4d = convert_input_2_numpy(input_4d_nifti)
output_numpy_4d = np.zeros_like(input_numpy_4d)
input_numpy_ktrans = convert_input_2_numpy(input_ktrans)
input_numpy_ve = convert_input_2_numpy(input_ve)
baseline_numpy = np.mean(input_numpy_4d[..., 0:7], axis=3)
scan_time_seconds = 307.2
time_interval_seconds = float((scan_time_seconds) / input_numpy_4d.shape[-1])
time_interval_minutes = time_interval_seconds/60
time_series = np.arange(0, input_numpy_4d.shape[-1]) / (60 / time_interval_seconds)
injection_start_time_seconds=38.4
T1_tissue=1000
T1_blood=1440
TR=3.8
flip_angle_degrees=25
relaxivity=.0045
hematocrit=.45
if input_aif == '':
population_AIF = parker_model_AIF(scan_time_seconds, injection_start_time_seconds, time_interval_seconds, input_numpy_4d)
concentration_AIF = population_AIF
else:
print 'extracting AIF...'
AIF_label_numpy = convert_input_2_numpy(input_aif)
AIF = generate_AIF(scan_time_seconds, injection_start_time_seconds, time_interval_seconds, input_numpy_4d, AIF_label_numpy)
concentration_AIF = convert_intensity_to_concentration(AIF, T1_tissue, TR, flip_angle_degrees, injection_start_time_seconds, relaxivity, time_interval_seconds, hematocrit, T1_blood=T1_blood)
for index in np.ndindex(input_numpy_ktrans.shape):
output_numpy_4d[index] = np.array(estimate_concentration([input_numpy_ktrans[index],input_numpy_ve[index]], concentration_AIF, time_interval_minutes))
# Or load presaved..
# output_numpy_4d = nifti_2_numpy('DCE_MRI_Phantom_Regenerated_Concentrations.nii.gz')
save_numpy_2_nifti(output_numpy_4d, input_4d_nifti, output_filepath + '_Regenerated_Concentrations.nii.gz')
output_numpy_4d = revert_concentration_to_intensity(data_numpy=output_numpy_4d, reference_data_numpy=input_numpy_4d, T1_tissue=T1_tissue, TR=TR, flip_angle_degrees=flip_angle_degrees, injection_start_time_seconds=injection_start_time_seconds, relaxivity=relaxivity, time_interval_seconds=time_interval_seconds, hematocrit=hematocrit, T1_blood=0, T1_map = [])
save_numpy_2_nifti(output_numpy_4d, input_4d_nifti, output_filepath + '_Regenerated_Signal.nii.gz')
return
def __init__(self,
n_samples=1000,
max_seq_len=60,
min_seq_len=25,
ktrans_range=[.001, 2],
ve_range=[0.01, .99],
gaussian_noise=[0, 0],
T1_range=[1000, 1000],
TR_range=[5, 5],
flip_angle_degrees_range=[30, 30],
relaxivity_range=[.0045, .0045],
hematocrit_range=[.45, .45],
sequence_length_range=[50, 50],
time_interval_seconds_range=[2, 2],
injection_start_time_seconds_range=[10, 10],
T1_blood_range=[1440, 1440],
baseline_intensity=[100, 100]):
ktrans_low_range = [.001, .3]
self.data = []
self.labels = []
self.seqlen = []
for i in range(n_samples):
# Random sequence length
seq_len = np.random.random_integers(*sequence_length_range)
# Monitor sequence length for TensorFlow dynamic calculation
self.seqlen.append(seq_len)
# Add a random or linear int sequence (50% prob)
if random.random() < .5 or True:
injection_start_time_seconds = np.random.uniform(
*injection_start_time_seconds_range)
time_interval_seconds = np.random.uniform(
*time_interval_seconds_range)
time_interval_minutes = time_interval_seconds / 60
scan_time_seconds = seq_len * time_interval_seconds
# Adjust for Unrealistically late injection time. Do this in a one liner later.
while injection_start_time_seconds > .8 * scan_time_seconds:
injection_start_time_seconds = np.random.uniform(
*injection_start_time_seconds_range)
AIF = parker_model_AIF(scan_time_seconds,
injection_start_time_seconds,
time_interval_seconds,
timepoints=seq_len)
ktrans = np.random.uniform(*ktrans_range)
ve = np.random.uniform(*ve_range)
Concentration = np.array(
estimate_concentration([ktrans, ve], AIF,
time_interval_minutes))
Intensity = revert_concentration_to_intensity(
data_numpy=Concentration,
reference_data_numpy=[],
T1_tissue=np.random.uniform(*T1_range),
TR=np.random.uniform(*TR_range),
flip_angle_degrees=np.random.uniform(
*flip_angle_degrees_range),
injection_start_time_seconds=injection_start_time_seconds,
relaxivity=np.random.uniform(*relaxivity_range),
time_interval_seconds=time_interval_seconds,
hematocrit=np.random.uniform(*hematocrit_range),
T1_blood=0,
T1_map=None,
static_baseline=np.random.uniform(
*baseline_intensity)).tolist()
s = Intensity
# Normalize
# Intensity = (Intensity - np.mean(Intensity)) / np.std(Intensity)
s += [0. for i in range(max_seq_len - seq_len)]
s = [[i] for i in s]
self.data.append(s)
self.labels.append([ktrans, ve])
self.batch_id = 0
def __init__(self,
n_samples=1000,
max_seq_len=65,
min_seq_len=56,
patch_x=5,
patch_y=5,
ktrans_range=[.001, 2],
ve_range=[0.01, .99],
gaussian_noise=[0, 0],
T1_range=[900, 1500],
TR_range=[3, 6],
flip_angle_degrees_range=[15, 35],
relaxivity_range=[.0045, .0045],
hematocrit_range=[.45, .45],
sequence_length_range=[65, 65],
time_interval_seconds_range=[1.5, 6],
injection_start_time_seconds_range=[8, 20],
T1_blood_range=[1440, 1440],
baseline_intensity=[20, 300]):
self.data = []
self.labels = []
self.seqlen = []
for i in range(n_samples):
# Random sequence length
seq_len = np.random.random_integers(*sequence_length_range)
# Monitor sequence length for TensorFlow dynamic calculation
self.seqlen.append(seq_len)
# Add a random or linear int sequence (50% prob)
if random.random() < .5 or True:
injection_start_time_seconds = np.random.uniform(
*injection_start_time_seconds_range)
time_interval_seconds = np.random.uniform(
*time_interval_seconds_range)
time_interval_minutes = time_interval_seconds / 60
scan_time_seconds = seq_len * time_interval_seconds
# Adjust for Unrealistically late injection time. Do this in a one liner later.
while injection_start_time_seconds > .8 * scan_time_seconds:
injection_start_time_seconds = np.random.uniform(
*injection_start_time_seconds_range)
AIF = parker_model_AIF(scan_time_seconds,
injection_start_time_seconds,
time_interval_seconds,
timepoints=seq_len)
true_ktrans = np.random.uniform(*ktrans_range)
true_ve = np.random.uniform(*ve_range)
ktrans = true_ktrans + np.zeros((patch_x, patch_y))
ve = true_ve + np.zeros((patch_x, patch_y))
Concentration = np.array(
estimate_concentration_general([ktrans, ve], AIF,
time_interval_minutes))
Concentration = Concentration * np.abs(
np.random.normal(1, .25, Concentration.shape))
Intensity = revert_concentration_to_intensity(
data_numpy=Concentration,
reference_data_numpy=[],
T1_tissue=np.random.uniform(*T1_range),
TR=np.random.uniform(*TR_range),
flip_angle_degrees=np.random.uniform(
*flip_angle_degrees_range),
injection_start_time_seconds=injection_start_time_seconds,
relaxivity=np.random.uniform(*relaxivity_range),
time_interval_seconds=time_interval_seconds,
hematocrit=np.random.uniform(*hematocrit_range),
T1_blood=0,
T1_map=None,
static_baseline=np.random.uniform(
*baseline_intensity)).tolist()
s = np.expand_dims(Intensity, -1)
self.data.append(s)
self.labels.append([true_ktrans, true_ve])
self.batch_id = 0
def __init__(self,
n_samples=1000,
max_seq_len=50,
min_seq_len=3,
max_value=1000,
ktrans_range=[.001, 2],
ve_range=[0.001, .95],
gaussian_noise=[0, 0],
T1_range=[1000, 1000],
TR_range=[5, 5],
flip_angle_degrees_range=[30, 30],
relaxivity_range=[.0045, .0045],
hematocrit_range=[.45, .45],
sequence_length_range=[50, 50],
time_interval_seconds_range=[2, 2],
injection_start_time_seconds_range=[10, 10],
T1_blood_range=[1440, 1440],
baseline_intensity=[100, 100]):
ktrans_low_range = [.001, .3]
self.data = []
self.labels = []
self.seqlen = []
for i in range(n_samples):
# Random sequence length
seq_len = np.random.random_integers(*sequence_length_range)
# Monitor sequence length for TensorFlow dynamic calculation
self.seqlen.append(seq_len)
# Add a random or linear int sequence (50% prob)
if random.random() < .5 or True:
injection_start_time_seconds = np.random.uniform(
*injection_start_time_seconds_range)
time_interval_seconds = np.random.uniform(
*time_interval_seconds_range)
time_interval_minutes = time_interval_seconds / 60
scan_time_seconds = seq_len * time_interval_seconds
while injection_start_time_seconds > .8 * scan_time_seconds:
injection_start_time_seconds = np.random.uniform(
*injection_start_time_seconds_range)
AIF = parker_model_AIF(scan_time_seconds,
injection_start_time_seconds,
time_interval_seconds,
timepoints=seq_len)
ktrans = np.random.uniform(*ktrans_range)
ve = np.random.uniform(*ve_range)
Concentration = np.array(
estimate_concentration([ktrans, ve], AIF,
time_interval_minutes))
Intensity = revert_concentration_to_intensity(
data_numpy=Concentration,
reference_data_numpy=[],
T1_tissue=np.random.uniform(*T1_range),
TR=np.random.uniform(*TR_range),
flip_angle_degrees=np.random.uniform(
*flip_angle_degrees_range),
injection_start_time_seconds=injection_start_time_seconds,
relaxivity=np.random.uniform(*relaxivity_range),
time_interval_seconds=time_interval_seconds,
hematocrit=np.random.uniform(*hematocrit_range),
T1_blood=0,
T1_map=[],
static_baseline=np.random.uniform(
*baseline_intensity)).tolist()
s = Intensity
# Intensity = (Intensity - np.mean(Intensity)) / np.std(Intensity)
# s = []
# s += [[value] for value in Intensity]
s += [0. for i in range(max_seq_len - seq_len)]
self.data.append(s)
self.labels.append([ktrans])
# else:
# # Generate a random sequence
# s = [[np.random.uniform(baseline_intensity[0]*2, baseline_intensity[0]*3)]
# for i in range(seq_len)]
# # Pad sequence for dimension consistency
# s += [[0.] for i in range(max_seq_len - seq_len)]
# self.data.append(s)
# self.labels.append([0., 1.])
# else:
# injection_start_time_seconds = np.random.uniform(*injection_start_time_seconds_range)
# time_interval_seconds = np.random.uniform(*time_interval_seconds_range)
# time_interval_minutes = time_interval_seconds/60
# scan_time_seconds = seq_len * time_interval_seconds
# while injection_start_time_seconds > .8*scan_time_seconds:
# injection_start_time_seconds = np.random.uniform(*injection_start_time_seconds_range)
# AIF = parker_model_AIF(scan_time_seconds, injection_start_time_seconds, time_interval_seconds, timepoints=seq_len)
# Concentration = np.array(estimate_concentration([np.random.uniform(*ktrans_low_range),np.random.uniform(*ve_range)], AIF, time_interval_minutes))
# Intensity = revert_concentration_to_intensity(data_numpy=Concentration, reference_data_numpy=[], T1_tissue=np.random.uniform(*T1_range), TR=np.random.uniform(*TR_range), flip_angle_degrees=np.random.uniform(*flip_angle_degrees_range), injection_start_time_seconds=injection_start_time_seconds, relaxivity=np.random.uniform(*relaxivity_range), time_interval_seconds=time_interval_seconds, hematocrit=np.random.uniform(*hematocrit_range), T1_blood=0, T1_map = [], static_baseline=np.random.uniform(*baseline_intensity)).tolist()
# s = []
# Intensity = Intensity - np.mean(Intensity) / np.std(Intensity)
# s += [[value] for value in Intensity]
# s += [[0.] for i in range(max_seq_len - seq_len)]
# self.data.append(s)
# self.labels.append([0., 1.])
self.batch_id = 0