for i in range(ndirs):
im_dc[i, :, :] = tf.ifft2c(data_dc[i, :, :], ph_b1[i, :, :])
# A way to quickly plot the different direction cases
if 0:
for i in range(ndirs):
ims = [
im_b0_avg, im_b1_full[i, :, :], im_b1_scan[i, :, :], im_dc[i, :, :]
]
titles = [
'b0 Average', 'im_full dir' + str(i), 'im_scan dir' + str(i),
'im_dc dir' + str(i)
]
vis.figSubplots(ims, clim=(minval, maxval), titles=titles)
saveFig.save(
'/micehome/asalerno/Documents/pyDirectionCompSense/brainData/DTI/26apr16/33per/imdc/im_b1_comps_dir'
+ str(i))
im_sp = im_dc.copy().reshape(N)
# Optimization algortihm -- this is where everything culminates together
#M, dIM, Ause, inds = dirInfo
for i in range(len(TV)):
args = (N, TV[i], XFM[i], data_b1, k, strtag, ph_b1, dirWeight, dirs,
dirInfo, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun,
im_dc,
args=args,
method=method,
jac=derivative_fun,
def runCSAlgorithm(fromfid=False,
filename='/home/asalerno/Documents/pyDirectionCompSense/brainData/P14/data/fullySampledBrain.npy',
sliceChoice=150,
strtag = ['','spatial', 'spatial'],
xtol = [1e-2, 1e-3, 5e-4, 5e-4],
TV = [0.01, 0.005, 0.002, 0.001],
XFM = [0.01,.005, 0.002, 0.001],
dirWeight=0,
pctg=0.25,
radius=0.2,
P=2,
pft=False,
ext=0.5,
wavelet='db4',
mode='per',
method='CG',
ItnLim=30,
lineSearchItnLim=30,
alpha_0=0.6,
c=0.6,
a=10.0,
kern =
np.array([[[ 0., 0., 0.],
[ 0., 0., 0.],
[ 0., 0., 0.]],
[[ 0., 0., 0.],
[ 0., -1., 0.],
[ 0., 1., 0.]],
[[ 0., 0., 0.],
[ 0., -1., 1.],
[ 0., 0., 0.]]]),
dirFile = None,
nmins = None,
dirs = None,
M = None,
dirInfo = [None]*4,
saveNpy=False,
saveNpyFile=None,
saveImsPng=False,
saveImsPngFile=None,
saveImDiffPng=False,
saveImDiffPngFile=None,
disp=False):
##import pdb; pdb.set_trace()
if fromfid==True:
inputdirectory=filename[0]
petable=filename[1]
fullImData = rff.getDataFromFID(petable,inputdirectory,2)[0,:,:,:]
fullImData = fullImData/np.max(abs(fullImData))
im = fullImData[:,:,sliceChoice]
else:
im = np.load(filename)[sliceChoice,:,:]
N = np.array(im.shape) # image Size
pdf = samp.genPDF(N[-2:], P, pctg, radius=radius, cyl=np.hstack([1, N[-2:]]), style='mult', pft=pft, ext=ext)
if pft:
print('Partial Fourier sampling method used')
k = samp.genSampling(pdf, 50, 2)[0].astype(int)
if len(N) == 2:
N = np.hstack([1, N])
k = k.reshape(N)
im = im.reshape(N)
elif (len(N) == 3) and ('dir' not in strtag):
k = k.reshape(np.hstack([1,N[-2:]])).repeat(N[0],0)
ph_ones = np.ones(N[-2:], complex)
ph_scan = np.zeros(N, complex)
data = np.zeros(N,complex)
im_scan = np.zeros(N,complex)
for i in range(N[0]):
k[i,:,:] = np.fft.fftshift(k[i,:,:])
data[i,:,:] = k[i,:,:]*tf.fft2c(im[i,:,:], ph=ph_ones)
# IMAGE from the "scanner data"
im_scan_wph = tf.ifft2c(data[i,:,:], ph=ph_ones)
ph_scan[i,:,:] = tf.matlab_style_gauss2D(im_scan_wph,shape=(5,5))
ph_scan[i,:,:] = np.exp(1j*ph_scan[i,:,:])
im_scan[i,:,:] = tf.ifft2c(data[i,:,:], ph=ph_scan[i,:,:])
#im_lr = samp.loRes(im,pctg)
# ------------------------------------------------------------------ #
# A quick way to look at the PSF of the sampling pattern that we use #
delta = np.zeros(N[-2:])
delta[int(N[-2]/2),int(N[-1]/2)] = 1
psf = tf.ifft2c(tf.fft2c(delta,ph_ones)*k,ph_ones)
# ------------------------------------------------------------------ #
## ------------------------------------------------------------------ #
## -- Currently broken - Need to figure out what's happening here. -- #
## ------------------------------------------------------------------ #
#if pft:
#for i in xrange(N[0]):
#dataHold = np.fft.fftshift(data[i,:,:])
#kHold = np.fft.fftshift(k[i,:,:])
#loc = 98
#for ix in xrange(N[-2]):
#for iy in xrange(loc,N[-1]):
#dataHold[-ix,-iy] = dataHold[ix,iy].conj()
#kHold[-ix,-iy] = kHold[ix,iy]
## ------------------------------------------------------------------ #
pdfDiv = pdf.copy()
pdfZeros = np.where(pdf==0)
pdfDiv[pdfZeros] = 1
#im_scan_imag = im_scan.imag
#im_scan = im_scan.real
N_im = N.copy()
hld, dims, dimOpt, dimLenOpt = tf.wt(im_scan[0].real,wavelet,mode)
N = np.hstack([N_im[0], hld.shape])
w_scan = np.zeros(N)
w_full = np.zeros(N)
im_dc = np.zeros(N_im)
w_dc = np.zeros(N)
for i in xrange(N[0]):
w_scan[i,:,:] = tf.wt(im_scan.real[i,:,:],wavelet,mode,dims,dimOpt,dimLenOpt)[0]
w_full[i,:,:] = tf.wt(abs(im[i,:,:]),wavelet,mode,dims,dimOpt,dimLenOpt)[0]
im_dc[i,:,:] = tf.ifft2c(data[i,:,:] / np.fft.ifftshift(pdfDiv), ph=ph_scan[i,:,:]).real.copy()
w_dc[i,:,:] = tf.wt(im_dc,wavelet,mode,dims,dimOpt,dimLenOpt)[0]
w_dc = w_dc.flatten()
im_sp = im_dc.copy().reshape(N_im)
minval = np.min(abs(im))
maxval = np.max(abs(im))
data = np.ascontiguousarray(data)
imdcs = [im_dc,np.zeros(N_im),np.ones(N_im),np.random.randn(np.prod(N_im)).reshape(N_im)]
imdcs[-1] = imdcs[-1] - np.min(imdcs[-1])
imdcs[-1] = imdcs[-1]/np.max(abs(imdcs[-1]))
mets = ['Density Corrected','Zeros','1/2''s','Gaussian Random Shift (0,1)']
wdcs = []
for i in range(len(imdcs)):
wdcs.append(tf.wt(imdcs[i][0],wavelet,mode,dims,dimOpt,dimLenOpt)[0].reshape(N))
ims = []
#print('Starting the CS Algorithm')
for kk in range(len(wdcs)):
w_dc = wdcs[kk]
print(mets[kk])
for i in range(len(TV)):
args = (N, N_im, dims, dimOpt, dimLenOpt, TV[i], XFM[i], data, k, strtag, ph_scan, kern, dirWeight, dirs, dirInfo, nmins, wavelet, mode, a)
w_result = opt.minimize(f, w_dc, args=args, method=method, jac=df,
options={'maxiter': ItnLim, 'lineSearchItnLim': lineSearchItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': alpha_0, 'c': c, 'xtol': xtol[i], 'TVWeight': TV[i], 'XFMWeight': XFM[i], 'N': N})
if np.any(np.isnan(w_result['x'])):
print('Some nan''s found. Dropping TV and XFM values')
elif w_result['status'] != 0:
print('TV and XFM values too high -- no solution found. Dropping...')
else:
w_dc = w_result['x']
w_res = w_dc.reshape(N)
im_res = np.zeros(N_im)
for i in xrange(N[0]):
im_res[i,:,:] = tf.iwt(w_res[i,:,:],wavelet,mode,dims,dimOpt,dimLenOpt)
ims.append(im_res)
if saveNpy:
if saveNpyFile is None:
np.save('./holdSave_im_res_' + str(int(pctg*100)) + 'p_all_SP',ims)
else:
np.save(saveNpyFile,ims)
if saveImsPng:
vis.figSubplots(ims,titles=mets,clims=(minval,maxval),colorbar=True)
if not disp:
if saveImsPngFile is None:
saveFig.save('./holdSave_ims_' + str(int(pctg*100)) + 'p_all_SP')
else:
saveFig.save(saveImsPngFile)
if saveImDiffPng:
imdiffs, clims = vis.imDiff(ims)
diffMets = ['DC-Zeros','DC-Ones','DC-Random','Zeros-Ones','Zeros-Random','Ones-Random']
vis.figSubplots(imdiffs,titles=diffMets,clims=clims,colorbar=True)
if not disp:
if saveImDiffPngFile is None:
saveFig.save('./holdSave_im_diffs_' + str(int(pctg*100)) + 'p_all_SP')
else:
saveFig.save(saveImDiffPngFile)
if disp:
plt.show()
for i in range(4):
clims.append([0, 1])
vis.figSubplots([
abs(im[0]), im_scan[0].real, im_scanComb[0].real, im_scanDir[0].real,
im_scanDirComb[0].real, k[0], kDir[0], kComb[0], kDirComb[0]
],
clims=clims,
titles=[
'Original', 'Var Dens', 'Var Dens Comb', 'Uni Samp',
'Uni Samp Comb', 'Var Dens k', 'Combo Dens k',
'Var Dens Shared k', 'Combo Dens Shared k'
],
colorbar=False)
#plt.show()
saveFig.save('tests/directionTests/phantomSmoothed_teamWithVarDens')
plt.figure(figsize=(24, 13.5))
plt.subplot(2, 2, 1)
samp.radialHistogram(k[0], rmax=1)
plt.title('Radial Histogram Var Dens')
plt.subplot(2, 2, 2)
samp.radialHistogram(kDir[0], rmax=1)
plt.title('Radial Histogram Combo Dens')
plt.subplot(2, 2, 3)
samp.radialHistogram(kComb[0], rmax=1)
plt.title('Radial Histogram Var Dens Shared')
plt.subplot(2, 2, 4)
samp.radialHistogram(kDirComb[0], rmax=1)
plt.title('Radial Histogram Combo Dens Shared')
saveFig.save('tests/directionTests/phantomSmoothed_radialHistograms')
# Now we add the k-space maps together in order to utilize the redundancies ot our advantage
data_dc = np.fft.fftshift(d.dir_dataSharing(k, np.fft.fftshift(data_b1, axes=(-2,-1)), dirs=dirs,
origDataSize=N[-2:]), axes=(-2,-1))
im_dc = np.zeros(data_dc.shape, dtype='complex')
# And make the images
for i in range(ndirs):
im_dc[i,:,:] = tf.ifft2c(data_dc[i,:,:],ph_b1[i,:,:])
# A way to quickly plot the different direction cases
if 0:
for i in range(ndirs):
ims = [im_b0_avg, im_b1_full[i,:,:], im_b1_scan[i,:,:], im_dc[i,:,:]]
titles = ['b0 Average', 'im_full dir' + str(i), 'im_scan dir' + str(i), 'im_dc dir' + str(i)]
vis.figSubplots(ims,clim=(minval,maxval),titles=titles)
saveFig.save('/micehome/asalerno/Documents/pyDirectionCompSense/brainData/DTI/26apr16/33per/imdc/im_b1_comps_dir'+str(i))
im_sp = im_dc.copy().reshape(N)
# Optimization algortihm -- this is where everything culminates together
#M, dIM, Ause, inds = dirInfo
for i in range(len(TV)):
args = (N, TV[i], XFM[i], data_b1, k, strtag, ph_b1, dirWeight, dirs, dirInfo, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'lineSearchItnLim': lineSearchItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': alpha_0, 'c': c, 'xtol': xtol[i], 'TVWeight': TV[i], 'XFMWeight': XFM[i], 'N': N})
if np.any(np.isnan(im_result['x'])):
print('Some nan''s found. Dropping TV and XFM values')
else:
elif imdcs == 'densCorr_Completed':
#im_dc = np.load('/home/asalerno/Documents/pyDirectionCompSense/brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/TV0.005_XFM0.005/0.25per_result_im_dc_densCorr.npy')
im_dc = im_res
TVWeight = 0.001
XFMWeight = 0.001
xtol = 1e-4
# Optimization algortihm -- this is where everything culminates together
a = 10.0
args = (N, TVWeight, XFMWeight, data, k, strtag, ph, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun, options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': alpha_0, 'c': 0.6, 'xtol': xtol, 'TVWeight': TVWeight, 'XFMWeight': XFMWeight, 'N': N})
im_res = im_result['x'].reshape(N)
plt.imshow(im_res,clim=(minval,maxval),interpolation='bilinear')
plt.title("Reconstructed Image with %d%% Sampling and im_dc==" % (pctg*100) + imdcs)
plt.show()
if imdcs == 'densCorr':
saveFig.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_firstRun_im_dc_" % (pctg*100) + imdcs)
np.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_firstRun_im_dc_" % (pctg*100) + imdcs,im_res)
elif imdcs == 'densCorr_Completed':
saveFig.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_secondRun_im_dc_TVXFM_0.001" % (pctg*100) + imdcs)
np.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_secondRun_im_dc_TVXFM_0.001" % (pctg*100) + imdcs,im_res)
#lrLoc = int(np.ceil((294-np.ceil(294/np.sqrt(1/pctg)))/2))
#im_lr = tf.fft2c(np.fft.fftshift(np.fft.fftshift(tf.ifft2c(im,np.ones(im.shape)))[lrLoc:-lrLoc,lrLoc:-lrLoc]),np.ones(im[lrLoc:-lrLoc,lrLoc:-lrLoc].shape))
#plt.imshow(abs(im_lr))
#plt.title('Low Resolution Equivalent with only %.2f%% of k-space Sampled' % pctg)
#saveFig.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/TV%.2f_XFM%.2f/%.2fper_result_im_lr" % (TVWeight, XFMWeight, pctg))
minval = np.min(im)
maxval = np.max(im)
# Primary first guess. What we're using for now. Density corrected
#im_dc = tf.ifft2c(data / np.fft.ifftshift(pdf), ph=ph).real.flatten().copy()
for imdcs in ['zeros','ones','densCorr','imFull']:
if imdcs == 'zeros':
im_dc = np.zeros(data.shape)
elif imdcs == 'ones':
im_dc = np.ones(data.shape)
elif imdcs == 'densCorr':
im_dc = tf.ifft2c(data / np.fft.ifftshift(pdf), ph=ph).real.flatten().copy()
elif imdcs == 'imFull':
im_dc = im
# Optimization algortihm -- this is where everything culminates together
a = 10.0
args = (N, TVWeight, XFMWeight, data, k, strtag, ph, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': 0.1, 'c': 0.6, 'xtol': 5e-3, 'TVWeight': TVWeight, 'XFMWeight': XFMWeight, 'N': N})
im_res = im_result['x'].reshape(N)
plt.imshow(abs(im_res),clim=(minval,maxval))
plt.title("Reconstructed Image with %.2f%% Sampling and im_dc==" % pctg + imdcs)
saveFig.save("sheppLoganData/TV%s_XFM%s/%.2fper_result_im_dc_" % (float('%.1g' % TVWeight), float('%.1g' % XFMWeight) , pctg) + imdcs)
np.save("sheppLoganData/TV%s_XFM%s/%.2fper_result_im_dc_" % (float('%.1g' % TVWeight), float('%.1g' % XFMWeight) , pctg) + imdcs + ".npy" ,im_res)
lrLoc = int(np.ceil((N[0]-np.ceil(N[0]/np.sqrt(1/pctg)))/2))
im_lr = tf.fft2c(np.fft.fftshift(np.fft.fftshift(tf.ifft2c(im,np.ones(im.shape)))[lrLoc:-lrLoc,lrLoc:-lrLoc]),np.ones(im[lrLoc:-lrLoc,lrLoc:-lrLoc].shape))
plt.imshow(abs(im_lr))
plt.title('Low Resolution Equivalent with only %.2f%% of k-space Sampled' % pctg)
saveFig.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/TV%.2f_XFM%.2f/%.2fper_result_im_lr" % (TVWeight, XFMWeight, pctg))
np.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/TV%.2f_XFM%.2f/%.2fper_result_im_lr" % (TVWeight, XFMWeight, pctg) + ".npy", im_lr)
'gtol': 0.01,
'disp': 1,
'alpha_0': 0.1,
'c': 0.6,
'xtol': 5e-3,
'TVWeight': TVWeight,
'XFMWeight': XFMWeight,
'N': N
})
im_res = im_result['x'].reshape(N)
plt.imshow(abs(im_res), clim=(minval, maxval))
plt.title(
"Reconstructed Image with %.2f%% Sampling and im_dc==" %
pctg + imdcs)
saveFig.save(
"sheppLoganData/TV%s_XFM%s/%.2fper_result_im_dc_" %
(float('%.1g' % TVWeight), float('%.1g' % XFMWeight),
pctg) + imdcs)
np.save(
"sheppLoganData/TV%s_XFM%s/%.2fper_result_im_dc_" %
(float('%.1g' % TVWeight), float('%.1g' % XFMWeight), pctg)
+ imdcs + ".npy", im_res)
lrLoc = int(np.ceil((N[0] - np.ceil(N[0] / np.sqrt(1 / pctg))) / 2))
im_lr = tf.fft2c(
np.fft.fftshift(
np.fft.fftshift(tf.ifft2c(im, np.ones(im.shape)))[lrLoc:-lrLoc,
lrLoc:-lrLoc]),
np.ones(im[lrLoc:-lrLoc, lrLoc:-lrLoc].shape))
plt.imshow(abs(im_lr))
plt.title('Low Resolution Equivalent with only %.2f%% of k-space Sampled' %
pctg)
for i in range(4):
clims.append([0,1])
vis.figSubplots([abs(im[0]),
im_scan[0].real,
im_scanComb[0].real,
im_scanDir[0].real,
im_scanDirComb[0].real,
k[0],
kDir[0],
kComb[0],
kDirComb[0]],
clims=clims,
titles=['Original','Var Dens','Var Dens Comb','Uni Samp','Uni Samp Comb','Var Dens k','Combo Dens k','Var Dens Shared k','Combo Dens Shared k'],colorbar=False)
#plt.show()
saveFig.save('tests/directionTests/phantomSmoothed_teamWithVarDens')
plt.figure(figsize=(24,13.5))
plt.subplot(2,2,1)
samp.radialHistogram(k[0],rmax=1)
plt.title('Radial Histogram Var Dens')
plt.subplot(2,2,2)
samp.radialHistogram(kDir[0],rmax=1)
plt.title('Radial Histogram Combo Dens')
plt.subplot(2,2,3)
samp.radialHistogram(kComb[0],rmax=1)
plt.title('Radial Histogram Var Dens Shared')
plt.subplot(2,2,4)
samp.radialHistogram(kDirComb[0],rmax=1)
plt.title('Radial Histogram Combo Dens Shared')
'c': 0.6,
'xtol': xtol,
'TVWeight': TVWeight,
'XFMWeight': XFMWeight,
'N': N
})
im_res = im_result['x'].reshape(N)
plt.imshow(im_res, clim=(minval, maxval), interpolation='bilinear')
plt.title("Reconstructed Image with %d%% Sampling and im_dc==" %
(pctg * 100) + imdcs)
plt.show()
if imdcs == 'densCorr':
saveFig.save(
"brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_firstRun_im_dc_"
% (pctg * 100) + imdcs)
np.save(
"brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_firstRun_im_dc_"
% (pctg * 100) + imdcs, im_res)
elif imdcs == 'densCorr_Completed':
saveFig.save(
"brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_secondRun_im_dc_TVXFM_0.001"
% (pctg * 100) + imdcs)
np.save(
"brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/replicate/seed1000_%.2fper_secondRun_im_dc_TVXFM_0.001"
% (pctg * 100) + imdcs, im_res)
#lrLoc = int(np.ceil((294-np.ceil(294/np.sqrt(1/pctg)))/2))
#im_lr = tf.fft2c(np.fft.fftshift(np.fft.fftshift(tf.ifft2c(im,np.ones(im.shape)))[lrLoc:-lrLoc,lrLoc:-lrLoc]),np.ones(im[lrLoc:-lrLoc,lrLoc:-lrLoc].shape))
#plt.imshow(abs(im_lr))
def runCSAlgorithm(
fromfid=False,
filename='/home/asalerno/Documents/pyDirectionCompSense/brainData/P14/data/fullySampledBrain.npy',
sliceChoice=150,
strtag=['', 'spatial', 'spatial'],
xtol=[1e-2, 1e-3, 5e-4, 5e-4],
TV=[0.01, 0.005, 0.002, 0.001],
XFM=[0.01, .005, 0.002, 0.001],
dirWeight=0,
pctg=0.25,
radius=0.2,
P=2,
pft=False,
ext=0.5,
wavelet='db4',
mode='per',
method='CG',
ItnLim=30,
lineSearchItnLim=30,
alpha_0=0.6,
c=0.6,
a=10.0,
kern=np.array([[[0., 0., 0.], [0., 0., 0.], [0., 0., 0.]],
[[0., 0., 0.], [0., -1., 0.], [0., 1., 0.]],
[[0., 0., 0.], [0., -1., 1.], [0., 0., 0.]]]),
dirFile=None,
nmins=None,
dirs=None,
M=None,
dirInfo=[None] * 4,
saveNpy=False,
saveNpyFile=None,
saveImsPng=False,
saveImsPngFile=None,
saveImDiffPng=False,
saveImDiffPngFile=None,
disp=False):
##import pdb; pdb.set_trace()
if fromfid == True:
inputdirectory = filename[0]
petable = filename[1]
fullImData = rff.getDataFromFID(petable, inputdirectory, 2)[0, :, :, :]
fullImData = fullImData / np.max(abs(fullImData))
im = fullImData[:, :, sliceChoice]
else:
im = np.load(filename)[sliceChoice, :, :]
N = np.array(im.shape) # image Size
pdf = samp.genPDF(N[-2:],
P,
pctg,
radius=radius,
cyl=np.hstack([1, N[-2:]]),
style='mult',
pft=pft,
ext=ext)
if pft:
print('Partial Fourier sampling method used')
k = samp.genSampling(pdf, 50, 2)[0].astype(int)
if len(N) == 2:
N = np.hstack([1, N])
k = k.reshape(N)
im = im.reshape(N)
elif (len(N) == 3) and ('dir' not in strtag):
k = k.reshape(np.hstack([1, N[-2:]])).repeat(N[0], 0)
ph_ones = np.ones(N[-2:], complex)
ph_scan = np.zeros(N, complex)
data = np.zeros(N, complex)
im_scan = np.zeros(N, complex)
for i in range(N[0]):
k[i, :, :] = np.fft.fftshift(k[i, :, :])
data[i, :, :] = k[i, :, :] * tf.fft2c(im[i, :, :], ph=ph_ones)
# IMAGE from the "scanner data"
im_scan_wph = tf.ifft2c(data[i, :, :], ph=ph_ones)
ph_scan[i, :, :] = tf.matlab_style_gauss2D(im_scan_wph, shape=(5, 5))
ph_scan[i, :, :] = np.exp(1j * ph_scan[i, :, :])
im_scan[i, :, :] = tf.ifft2c(data[i, :, :], ph=ph_scan[i, :, :])
#im_lr = samp.loRes(im,pctg)
# ------------------------------------------------------------------ #
# A quick way to look at the PSF of the sampling pattern that we use #
delta = np.zeros(N[-2:])
delta[int(N[-2] / 2), int(N[-1] / 2)] = 1
psf = tf.ifft2c(tf.fft2c(delta, ph_ones) * k, ph_ones)
# ------------------------------------------------------------------ #
## ------------------------------------------------------------------ #
## -- Currently broken - Need to figure out what's happening here. -- #
## ------------------------------------------------------------------ #
#if pft:
#for i in xrange(N[0]):
#dataHold = np.fft.fftshift(data[i,:,:])
#kHold = np.fft.fftshift(k[i,:,:])
#loc = 98
#for ix in xrange(N[-2]):
#for iy in xrange(loc,N[-1]):
#dataHold[-ix,-iy] = dataHold[ix,iy].conj()
#kHold[-ix,-iy] = kHold[ix,iy]
## ------------------------------------------------------------------ #
pdfDiv = pdf.copy()
pdfZeros = np.where(pdf == 0)
pdfDiv[pdfZeros] = 1
#im_scan_imag = im_scan.imag
#im_scan = im_scan.real
N_im = N.copy()
hld, dims, dimOpt, dimLenOpt = tf.wt(im_scan[0].real, wavelet, mode)
N = np.hstack([N_im[0], hld.shape])
w_scan = np.zeros(N)
w_full = np.zeros(N)
im_dc = np.zeros(N_im)
w_dc = np.zeros(N)
for i in xrange(N[0]):
w_scan[i, :, :] = tf.wt(im_scan.real[i, :, :], wavelet, mode, dims,
dimOpt, dimLenOpt)[0]
w_full[i, :, :] = tf.wt(abs(im[i, :, :]), wavelet, mode, dims, dimOpt,
dimLenOpt)[0]
im_dc[i, :, :] = tf.ifft2c(data[i, :, :] / np.fft.ifftshift(pdfDiv),
ph=ph_scan[i, :, :]).real.copy()
w_dc[i, :, :] = tf.wt(im_dc, wavelet, mode, dims, dimOpt, dimLenOpt)[0]
w_dc = w_dc.flatten()
im_sp = im_dc.copy().reshape(N_im)
minval = np.min(abs(im))
maxval = np.max(abs(im))
data = np.ascontiguousarray(data)
imdcs = [
im_dc,
np.zeros(N_im),
np.ones(N_im),
np.random.randn(np.prod(N_im)).reshape(N_im)
]
imdcs[-1] = imdcs[-1] - np.min(imdcs[-1])
imdcs[-1] = imdcs[-1] / np.max(abs(imdcs[-1]))
mets = [
'Density Corrected', 'Zeros', '1/2'
's', 'Gaussian Random Shift (0,1)'
]
wdcs = []
for i in range(len(imdcs)):
wdcs.append(
tf.wt(imdcs[i][0], wavelet, mode, dims, dimOpt,
dimLenOpt)[0].reshape(N))
ims = []
#print('Starting the CS Algorithm')
for kk in range(len(wdcs)):
w_dc = wdcs[kk]
print(mets[kk])
for i in range(len(TV)):
args = (N, N_im, dims, dimOpt, dimLenOpt, TV[i], XFM[i], data, k,
strtag, ph_scan, kern, dirWeight, dirs, dirInfo, nmins,
wavelet, mode, a)
w_result = opt.minimize(f,
w_dc,
args=args,
method=method,
jac=df,
options={
'maxiter': ItnLim,
'lineSearchItnLim': lineSearchItnLim,
'gtol': 0.01,
'disp': 1,
'alpha_0': alpha_0,
'c': c,
'xtol': xtol[i],
'TVWeight': TV[i],
'XFMWeight': XFM[i],
'N': N
})
if np.any(np.isnan(w_result['x'])):
print('Some nan' 's found. Dropping TV and XFM values')
elif w_result['status'] != 0:
print(
'TV and XFM values too high -- no solution found. Dropping...'
)
else:
w_dc = w_result['x']
w_res = w_dc.reshape(N)
im_res = np.zeros(N_im)
for i in xrange(N[0]):
im_res[i, :, :] = tf.iwt(w_res[i, :, :], wavelet, mode, dims,
dimOpt, dimLenOpt)
ims.append(im_res)
if saveNpy:
if saveNpyFile is None:
np.save('./holdSave_im_res_' + str(int(pctg * 100)) + 'p_all_SP',
ims)
else:
np.save(saveNpyFile, ims)
if saveImsPng:
vis.figSubplots(ims,
titles=mets,
clims=(minval, maxval),
colorbar=True)
if not disp:
if saveImsPngFile is None:
saveFig.save('./holdSave_ims_' + str(int(pctg * 100)) +
'p_all_SP')
else:
saveFig.save(saveImsPngFile)
if saveImDiffPng:
imdiffs, clims = vis.imDiff(ims)
diffMets = [
'DC-Zeros', 'DC-Ones', 'DC-Random', 'Zeros-Ones', 'Zeros-Random',
'Ones-Random'
]
vis.figSubplots(imdiffs, titles=diffMets, clims=clims, colorbar=True)
if not disp:
if saveImDiffPngFile is None:
saveFig.save('./holdSave_im_diffs_' + str(int(pctg * 100)) +
'p_all_SP')
else:
saveFig.save(saveImDiffPngFile)
if disp:
plt.show()
clim.append(np.min(im),np.max(im))
for j in range(len(TV)):
for k in range(len(XFM)):
for m in range(len(lam_trust)):
imRMSE[i,j,k,m,:,:] = np.load('temp/' + str(int(100*pctg)) + '_slice_' + str(SL[i]) + '_TV_' + str(TV[j]) + '_XFM_' + str(XFM[k]) + '_lamTrust_' + str(int(100*lam_trust[m])) + '.npy')
rmses[i,j,k,m] = rmse(imRMSE[i,j,k,m,:,:],im)
titles = []
for j in range(len(TV)):
for k in range(len(XFM)):
titles.append('TV = ' + str(TV[j]) + ' XFM = ' + str(XFM[k]))
for i in range(len(SL)):
for m in range(len(lam_trust)):
labs = ['Slice ' + str(SL[i]) + ' with $\lambda_{trust}$ = ' + str(lam_trust[m]), 'TV', 'XFM']
vis.figSubplots(np.concatenate(imRMSE[i,:,:,m,:,:],axis=0),titles=titles,labs=labs)
saveFig.save('/hpf/largeprojects/MICe/asalerno/pyDirectionCompSense/tests/rmseTests/ims_sl_' + str(SL[i]) + '_lamTrust_' + str(lam_trust[m]) + '_allTVXFM')
fig, ax = plt.subplots()
img = ax.imshow(rmses[i,:,:,m],cmap='jet',interpolation='None')
ax.set_title('RMSE: Slice ' + str(SL[i]) + ' $\lambda_{trust}$ = ' + str(lam_trust[m]))
strTV = ['']
strXFM = ['']
[strTV.append(TV[q]) for q in range(len(TV))]
[strXFM.append(XFM[q]) for q in range(len(XFM))]
ax.set_xticklabels(strTV)
ax.set_yticklabels(strXFM)
fig.colorbar(img)
saveFig.save('/hpf/largeprojects/MICe/asalerno/pyDirectionCompSense/tests/rmseTests/rmses_sl_' + str(SL[i]) + '_lamTrust_' + str(lam_trust[m]) + '_allTVXFM')
tvStps.append(TV[i])
w_res = w_dc.reshape(N)
im_res = np.zeros(N_im)
for i in xrange(N[0]):
im_res[i,:,:] = tf.iwt(w_res[i,:,:],wavelet,mode,dims,dimOpt,dimLenOpt)
ims.append(im_res)
data_res = tf.fft2c(im_res,ph_scan)
data_diff = np.fft.fftshift(data_res-data)*k
plt.imshow(abs(data_diff)[0]); plt.colorbar();
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/phantomKernTests/noise_' + str(int(100*pctg)) + '_ksp_diff_abs')
plt.imshow(data_diff.real[0]); plt.colorbar();
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/phantomKernTests/noise_' + str(int(100*pctg)) + '_ksp_diff_real')
plt.imshow(data_diff.imag[0]); plt.colorbar();
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/phantomKernTests/noise_' + str(int(100*pctg)) + '_ksp_diff_imag')
im_stps = np.zeros([len(stps), N_im[-2], N_im[-1]])
gtv = np.zeros([len(stps), N_im[-2], N_im[-1]])
gxfm = np.zeros([len(stps), N_im[-2], N_im[-1]])
gdc = np.zeros([len(stps), N_im[-2], N_im[-1]])
for jj in range(len(stps)):
im_stps[jj,:,:] = tf.iwt(stps[jj].reshape(N[-2:]),wavelet,mode,dims,dimOpt,dimLenOpt)
gtv[jj,:,:] = grads.gTV(im_stps[jj,:,:].reshape(N_im),N_im,strtag, kern, 0, a=a)
gxfm[jj,:,:] = tf.iwt(grads.gXFM(stps[jj].reshape(N[-2:]),a=a),wavelet,mode,dims,dimOpt,dimLenOpt)
gdc[jj,:,:] = grads.gDataCons(im_stps[jj,:,:], N_im, ph_scan, data, k)
plt.imshow(imgsData[pc,TV,XFM,0,:,:],clim=[minval,maxval])
ax2.set_title('Recon %.2f%% samp imdc=zeros' % (pctg*100))
ax3 = fig.add_subplot(2,3,3)
plt.imshow(imgsData[pc,TV,XFM,1,:,:],clim=(minval,maxval))
ax3.set_title('Recon %.2f%% samp imdc=ones' % (pctg*100))
ax4 = fig.add_subplot(2,3,4)
plt.imshow(abs(im_lr))
ax4.set_title('Low Res with %.2f%% of Data' % (pctg*100))
ax5 = fig.add_subplot(2,3,5)
plt.imshow(imgsData[pc,TV,XFM,2,:,:],clim=(minval,maxval))
ax5.set_title('Recon %.2f%% samp imdc=densCorr' % (pctg*100))
ax6 = fig.add_subplot(2,3,6)
plt.imshow(imgsData[pc,TV,XFM,3,:,:],clim=(minval,maxval))
ax6.set_title('Recon %.2f%% samp imdc=imFull' % (pctg*100))
saveFig.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/TV%s_XFM%s/%.2fper_im_lr_edit_comparison" % (float('%.1g' % TVWeight), float('%.1g' % XFMWeight), pctg))
XFM=-1
TV=-1
# Rerunning data comparison
pctg = 0.25
im = np.load('/home/asalerno/Documents/pyDirectionCompSense/brainData/exercise_irradiation/bruker_data/running_C/P14/fullySampledBrain.npy')
N = im.shape
minval = np.min(im)
maxval = np.max(im)
lrLoc = int(np.ceil((N[0]-np.ceil(N[0]/np.sqrt(1/pctg)))/2))
im_lr = tf.fft2c(np.fft.fftshift(np.fft.fftshift(tf.ifft2c(im,np.ones(im.shape)))[lrLoc:-lrLoc,lrLoc:-lrLoc]),np.ones(im[lrLoc:-lrLoc,lrLoc:-lrLoc].shape))
#----------------#
# Step 1 #
#----------------#
args = (N, TV1, XFM1, data, k, strtag, ph, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': 0.1, 'c': 0.6, 'xtol': xtol1, 'TVWeight': TV1, 'XFMWeight': XFM1, 'N': N})
im_dc = im_result['x'].reshape(N)
#----------------#
# Step 2 #
#----------------#
args = (N, TV2, XFM2, data, k, strtag, ph, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': 0.1, 'c': 0.6, 'xtol': xtol2, 'TVWeight': TV2, 'XFMWeight': XFM2, 'N': N})
im_dc = im_result['x'].reshape(N)
#----------------#
# Step 3 #
#----------------#
args = (N, TV3, XFM3, data, k, strtag, ph, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': 0.1, 'c': 0.6, 'xtol': xtol3, 'TVWeight': TV3, 'XFMWeight': XFM3, 'N': N})
im_res = im_result['x'].reshape(N)
plt.imshow(abs(im_res),clim=(minval,maxval),interpolation='bilinear')
plt.title("25% Data -- P=" + str(P) + " -- 3 Rounds")
plt.xlabel('TV1=XFM1=' + str(TV1) + ' TV2=XFM2=' + str(TV2) + ' TV3=XFM3=' + str(TV3))
saveFig.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/P_analysis/1_overr_p/seed" + str(int(seeds[sdcnt])) + '/3rounds_P_' + str(P) + '_ph_1')
np.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/P_analysis/1_overr_p/seed" + str(int(seeds[sdcnt])) + "/3rounds_P_" + str(P) + '_ph_1.npy',im_res)
gtv[jj, :, :] = grads.gTV(im_stps[jj, :, :].reshape(N_im),
N_im,
strtag,
kern,
0,
a=a)
gxfm[jj, :, :] = tf.iwt(grads.gXFM(stps[jj].reshape(N[-2:]), a=a), wavelet,
mode, dims, dimOpt, dimLenOpt)
gdc[jj, :, :] = grads.gDataCons(im_stps[jj, :, :], N_im, ph_scan, data, k)
for i in xrange(len(stps)):
plt.imshow(im_stps[i])
plt.colorbar()
plt.title('Step with Negative Values')
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/gradTests/' +
str(int(100 * pctg)) + '_TV_XFM_' + str(TV[i]) + '_0_a_' +
str(int(a)) + '_negs')
vis.figSubplots([im_stps[i], gtv[i], gxfm[i], gdc[i]],
titles=['Step', 'gTV', 'gXFM', 'gDC'],
clims=[(minval, maxval), (np.min(gtv[i]), np.max(gtv[i])),
(np.min(gxfm[i]), np.max(gxfm[i])),
(np.min(gdc[i]), np.max(gdc[i]))])
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/gradTests/' +
str(int(100 * pctg)) + '_TV_XFM_' + str(TV[i]) + '_0_a_' +
str(int(a)) + '_grads')
plt.imshow(TV[i] * gtv[i] + TV[i] * gxfm[i] + gdc[i])
plt.colorbar()
plt.title('Total Gradient')
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/gradTests/' +
str(int(100 * pctg)) + '_TV_XFM_' + str(TV[i]) + '_a_' +
str(int(a)) + '_gradTotal')
im_dc,
args=args,
method=method,
jac=derivative_fun,
options={
'maxiter': ItnLim,
'gtol': 0.01,
'disp': 1,
'alpha_0': 0.1,
'c': 0.6,
'xtol': xtol3,
'TVWeight': TV3,
'XFMWeight': XFM3,
'N': N
})
im_res = im_result['x'].reshape(N)
plt.imshow(abs(im_res),
clim=(minval, maxval),
interpolation='bilinear')
plt.title("25% Data -- P=" + str(P) + " -- 3 Rounds")
plt.xlabel('TV1=XFM1=' + str(TV1) + ' TV2=XFM2=' + str(TV2) +
' TV3=XFM3=' + str(TV3))
saveFig.save(
"brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/P_analysis/1_overr_p/seed"
+ str(int(seeds[sdcnt])) + '/3rounds_P_' + str(P) + '_ph_1')
np.save(
"brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/P_analysis/1_overr_p/seed"
+ str(int(seeds[sdcnt])) + "/3rounds_P_" + str(P) + '_ph_1.npy',
im_res)
cols = ('b','k','r','g','m','c','y','0.5','0.75',(0.25,0.25,0.75),(1,0.25,0.5))
fig=plt.figure()
strPs = map(str,Ps)
for P in Ps:
print(P)
cnt += 1
plt.plot(rads,pdfmult[cnt,147,:],color=cols[cnt])
#samp.pltSlice(pdfmult[cnt,:,:],sl=147,rads=rads,col=cols[cnt])
plt.legend(strPs)
plt.xlim(-np.sqrt(2),np.sqrt(2))
plt.ylim(0,1.1)
plt.xlabel('Radius')
plt.ylabel('PDF')
plt.title('PDF using 1/|r|^p')
saveFig.save('/micehome/asalerno/Documents/pyDirectionCompSense/brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/pdf/r_-p_method_compared')
cnt = -1
pleg=[]
for P in Ps:
print(P)
cnt += 1
if P>2:
pleg.append(str(P))
plt.plot(rads,pdfadd[cnt,147,:],color=cols[cnt])
#samp.pltSlice(pdfmult[cnt,:,:],sl=147,rads=rads,col=cols[cnt])
plt.legend(strPs)
plt.xlim(-np.sqrt(2),np.sqrt(2))
plt.ylim(0,1.1)
plt.xlabel('Radius')
plt.ylabel('PDF')
w_dc = w_result['x']
stps.append(w_dc)
tvStps.append(TV[i])
w_res = w_dc.reshape(N)
im_res = np.zeros(N_im)
for i in xrange(N[0]):
im_res[i,:,:] = tf.iwt(w_res[i,:,:],wavelet,mode,dims,dimOpt,dimLenOpt)
ims.append(im_res)
im_stps = np.zeros([len(stps), N_im[-2], N_im[-1]])
gtv = np.zeros([len(stps), N_im[-2], N_im[-1]])
gxfm = np.zeros([len(stps), N_im[-2], N_im[-1]])
gdc = np.zeros([len(stps), N_im[-2], N_im[-1]])
for jj in range(len(stps)):
im_stps[jj,:,:] = tf.iwt(stps[jj].reshape(N[-2:]),wavelet,mode,dims,dimOpt,dimLenOpt)
gtv[jj,:,:] = grads.gTV(im_stps[jj,:,:].reshape(N_im),N_im,strtag, kern, 0, a=a)
gxfm[jj,:,:] = tf.iwt(grads.gXFM(stps[jj].reshape(N[-2:]),a=a),wavelet,mode,dims,dimOpt,dimLenOpt)
gdc[jj,:,:] = grads.gDataCons(im_stps[jj,:,:], N_im, ph_scan, data, k)
for i in xrange(len(stps)):
vis.figSubplots([im_stps[i],gtv[i],gxfm[i],gdc[i]],titles=['Step','gTV','gXFM','gDC'])
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/gradTests/'+ str(int(100*pctg)) + '_TV_XFM_'+ str(TV[i]) +'_grads')
plt.imshow(TV[i]*gtv[i] + TV[i]*gxfm[i] + gdc[i])
plt.colorbar()
plt.title('Total Gradient')
saveFig.save('/home/asalerno/Documents/pyDirectionCompSense/gradTests/'+ str(int(100*pctg)) + '_TV_XFM_'+ str(TV[i]) +'_gradTotal')
#----------------#
args = (N, TV3, XFM3, data, k, strtag, ph_scan, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': c*alpha_k, 'c': c, 'xtol': xtol3, 'TVWeight': TV3, 'XFMWeight': XFM3, 'N': N})
im_dc = im_result['x'].reshape(N)
alpha_k = im_result['alpha_k']
#----------------#
# Step 4 #
#----------------#
args = (N, TV4, XFM4, data, k, strtag, ph_scan, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': c*alpha_k, 'c': c, 'xtol': xtol4, 'TVWeight': TV4, 'XFMWeight': XFM4, 'N': N})
im_dc = im_result['x'].reshape(N)
alpha_k = im_result['alpha_k']
#----------------#
# Step 5 #
#----------------#
args = (N, TV5, XFM5, data, k, strtag, ph_scan, dirWeight, dirs, M, nmins, wavelet, mode, a)
im_result = opt.minimize(optfun, im_dc, args=args, method=method, jac=derivative_fun,
options={'maxiter': ItnLim, 'gtol': 0.01, 'disp': 1, 'alpha_0': c*alpha_k, 'c': c, 'xtol': xtol5, 'TVWeight': TV5, 'XFMWeight': XFM5, 'N': N})
im_res = im_result['x'].reshape(N)
plt.imshow(abs(im_res),clim=(minval,maxval),interpolation='bilinear')
plt.title("25% Data -- P=" + str(P) + " -- 5 Rounds")
plt.xlabel('TV=XFM= [' + str(TV1) + ', ' + str(TV2) + ', ' + str(TV3) + ', ' + str(TV4) + ', ' + str(TV5) + ']')
#plt.show()
saveFig.save("brainData/P14/5rounds_P_" + str(P) + '_pctg_' + str(pctg) + '_sl_' + str(sliceChoice))
#np.save("brainData/exercise_irradiation/bruker_data/running_C/P14/rad_0.1/P_analysis/3rounds_P_" + str(P) + '.npy',im_res)
