def testMultimodalRigidTransformationMultiScale3D_ecqmmf(
betaGT, level, nclasses, lambdaParam, mu, maxIter, tolerance):
betaGTRads = np.array(betaGT, dtype=np.float64)
betaGTRads[0:3] = np.copy(np.pi * betaGTRads[0:3] / 180.0)
movingName = 'data/t2/t2_icbm_normal_1mm_pn0_rf0_peeled.nii.gz'
fixedName = 'data/t1/t1_icbm_normal_1mm_pn0_rf0_peeled.nii.gz'
print 'Loading data...'
moving = nib.load(movingName)
moving = moving.get_data().squeeze()
moving = moving.astype(np.float64)
fixed = nib.load(fixedName)
fixed = fixed.get_data().squeeze()
fixed = fixed.astype(np.float64)
moving = (moving - moving.min()) / (moving.max() - moving.min())
fixed = (fixed - fixed.min()) / (fixed.max() - fixed.min())
print 'Applying GT transform...'
fixed = rcommon.applyRigidTransformation3D(fixed, betaGTRads)
sh = moving.shape
#leftPyramid=[i for i in rcommon.pyramid_gaussian_3D(moving, level)]
#rightPyramid=[i for i in rcommon.pyramid_gaussian_3D(fixed, level)]
fixedImage = fixed[:, sh[1] // 2, :].copy()
movingImage = moving[:, sh[1] // 2, :].copy()
print 'Initializing registration...'
meansFixed, variancesFixed, meansMoving, variancesMoving, joint = initializeECQMMFRegistration(
fixedImage, movingImage, nclasses, lambdaParam, mu, maxIter, tolerance)
joint = np.array(joint)
negLogLikelihood = np.zeros_like(joint)
ecqmmf_reg.compute_registration_neg_log_likelihood_constant_models(
fixedImage, movingImage, joint, meansFixed, meansMoving,
variancesFixed, variancesMoving, negLogLikelihood)
#ecqmmf_reg.initialize_registration_maximum_likelihood_probs(negLogLikelihood, joint):
ecqmmf.initialize_normalized_likelihood(negLogLikelihood, joint)
bufferN = np.array(range(nclasses * nclasses))
bufferD = np.array(range(nclasses * nclasses))
for iter_count in range(maxIter):
print 'Iter:', iter_count, '/', maxIter
ecqmmf_reg.iterate_marginals(negLogLikelihood, joint, lambdaParam, mu,
bufferN, bufferD)
#ecqmmf_reg.integrate_registration_probabilistic_weighted_tensor_field_products(double[:,:,:] q, double[:,:] diff, double[:,:,:,:] probs, double[:] weights)
ecqmmf_reg.compute_registration_neg_log_likelihood_constant_models(
fixedImage, movingImage, joint, meansFixed, meansMoving,
variancesFixed, variancesMoving, negLogLikelihood)
#----plot joint probability maps---
print 'Plotting joint probability maps...'
plt.figure()
plt.title('Joint probability maps')
for i in range(nclasses):
for j in range(nclasses):
plt.subplot(nclasses, nclasses, 1 + i * nclasses + j)
plt.imshow(joint[:, :, i, j], cmap=plt.cm.gray)
plt.title("F=" + str(i) + ", M=" + str(j))
#----plot negLogLikelihood maps---
print 'Plotting negLogLikelihood maps...'
plt.figure()
plt.title('neg-log-likelihood maps')
for i in range(nclasses):
for j in range(nclasses):
plt.subplot(nclasses, nclasses, 1 + i * nclasses + j)
plt.imshow(negLogLikelihood[:, :, i, j], cmap=plt.cm.gray)
plt.title("F=" + str(i) + ", M=" + str(j))
def testMultimodalRigidTransformationMultiScale3D_ecqmmf(betaGT, level, nclasses, lambdaParam, mu, maxIter, tolerance):
betaGTRads=np.array(betaGT, dtype=np.float64)
betaGTRads[0:3]=np.copy(np.pi*betaGTRads[0:3]/180.0)
movingName='data/t2/t2_icbm_normal_1mm_pn0_rf0_peeled.nii.gz'
fixedName ='data/t1/t1_icbm_normal_1mm_pn0_rf0_peeled.nii.gz'
print 'Loading data...'
moving=nib.load(movingName)
moving=moving.get_data().squeeze()
moving=moving.astype(np.float64)
fixed=nib.load(fixedName)
fixed=fixed.get_data().squeeze()
fixed=fixed.astype(np.float64)
moving=(moving-moving.min())/(moving.max()-moving.min())
fixed=(fixed-fixed.min())/(fixed.max()-fixed.min())
print 'Applying GT transform...'
fixed=rcommon.applyRigidTransformation3D(fixed, betaGTRads)
sh=moving.shape
#leftPyramid=[i for i in rcommon.pyramid_gaussian_3D(moving, level)]
#rightPyramid=[i for i in rcommon.pyramid_gaussian_3D(fixed, level)]
fixedImage=fixed[:,sh[1]//2,:].copy()
movingImage=moving[:,sh[1]//2,:].copy()
print 'Initializing registration...'
meansFixed, variancesFixed, meansMoving, variancesMoving, joint=initializeECQMMFRegistration(fixedImage, movingImage, nclasses, lambdaParam, mu, maxIter, tolerance)
joint=np.array(joint)
negLogLikelihood=np.zeros_like(joint)
ecqmmf_reg.compute_registration_neg_log_likelihood_constant_models(fixedImage, movingImage, joint, meansFixed, meansMoving, variancesFixed, variancesMoving, negLogLikelihood)
#ecqmmf_reg.initialize_registration_maximum_likelihood_probs(negLogLikelihood, joint):
ecqmmf.initialize_normalized_likelihood(negLogLikelihood, joint)
bufferN=np.array(range(nclasses*nclasses))
bufferD=np.array(range(nclasses*nclasses))
for iter_count in range(maxIter):
print 'Iter:',iter_count,'/',maxIter
ecqmmf_reg.iterate_marginals(negLogLikelihood, joint, lambdaParam, mu, bufferN, bufferD)
#ecqmmf_reg.integrate_registration_probabilistic_weighted_tensor_field_products(double[:,:,:] q, double[:,:] diff, double[:,:,:,:] probs, double[:] weights)
ecqmmf_reg.compute_registration_neg_log_likelihood_constant_models(fixedImage, movingImage, joint, meansFixed, meansMoving, variancesFixed, variancesMoving, negLogLikelihood)
#----plot joint probability maps---
print 'Plotting joint probability maps...'
plt.figure()
plt.title('Joint probability maps')
for i in range(nclasses):
for j in range(nclasses):
plt.subplot(nclasses,nclasses,1+i*nclasses+j)
plt.imshow(joint[:,:,i,j], cmap=plt.cm.gray)
plt.title("F="+str(i)+", M="+str(j))
#----plot negLogLikelihood maps---
print 'Plotting negLogLikelihood maps...'
plt.figure()
plt.title('neg-log-likelihood maps')
for i in range(nclasses):
for j in range(nclasses):
plt.subplot(nclasses,nclasses,1+i*nclasses+j)
plt.imshow(negLogLikelihood[:,:,i,j], cmap=plt.cm.gray)
plt.title("F="+str(i)+", M="+str(j))
def estimateNewECQMMFMultimodalDeformationField2D(fixed, moving, nclasses, lambdaMeasureField, lambdaDisplacement, mu, maxOuterIter, maxInnerIter, tolerance, previousDisplacement=None):
sh=fixed.shape
X0,X1=np.mgrid[0:sh[0], 0:sh[1]]
displacement =np.empty(shape=(fixed.shape)+(2,), dtype=np.float64)
gradientField =np.empty(shape=(fixed.shape)+(2,), dtype=np.float64)
totalDisplacement=np.zeros(shape=(fixed.shape)+(2,), dtype=np.float64)
gradientField =np.empty(shape=(fixed.shape)+(2,), dtype=np.float64)
residuals =np.zeros_like(fixed)
warped=None
if(previousDisplacement!=None):
totalDisplacement[...]=previousDisplacement
warped=ndimage.map_coordinates(moving, [X0+totalDisplacement[...,0], X1+totalDisplacement[...,1]], prefilter=True)
else:
warped=moving
#run soft segmentation on the fixed image
meansFixed, variancesFixed=ecqmmf.initialize_constant_models(fixed, nclasses)
meansFixed=np.array(meansFixed)
variancesFixed=np.array(variancesFixed)
segFixed, meansFixed, variancesFixed, probsFixed=ecqmmf.ecqmmf(fixed, nclasses, lambdaMeasureField, mu, maxOuterIter, maxInnerIter, tolerance)
meansFixed=np.array(meansFixed)
probsFixed=np.array(probsFixed)
#run soft segmentation on the warped image
meansWarped, variancesWarped=ecqmmf.initialize_constant_models(warped, nclasses)
meansWarped=np.array(meansWarped)
variancesWarped=np.array(variancesWarped)
segWarped, meansWarped, variancesWarped, probsWarped=ecqmmf.ecqmmf(warped, nclasses, lambdaMeasureField, mu, maxOuterIter, maxInnerIter, tolerance)
meansWarped=np.array(meansWarped)
probsWarped=np.array(probsWarped)
#inicialize the joint models (solve assignment problem)
ecqmmf_reg.initialize_coupled_constant_models(probsFixed, probsWarped, meansWarped)
#start optimization
outerIter=0
negLogLikelihood=np.zeros_like(probsFixed)
while(outerIter<maxOuterIter):
outerIter+=1
print "Outer:", outerIter
if(outerIter>1):#avoid warping twice at the first iteration
warped=ndimage.map_coordinates(moving, [X0+totalDisplacement[...,0], X1+totalDisplacement[...,1]], prefilter=True)
movingMask=(moving>0)*1.0
warpedMask=ndimage.map_coordinates(movingMask, [X0+totalDisplacement[...,0], X1+totalDisplacement[...,1]], order=0, prefilter=False)
warpedMask=warpedMask.astype(np.int32)
#--- optimize the measure field and the intensity models ---
ecqmmf_reg.compute_registration_neg_log_likelihood_constant_models(fixed, warped, meansFixed, meansWarped, negLogLikelihood)
#ecqmmf.initialize_normalized_likelihood(negLogLikelihood, probsWarped)
ecqmmf.initialize_maximum_likelihood(negLogLikelihood, probsWarped);
innerIter=0
mse=0
while(innerIter<maxInnerIter):
innerIter+=1
print "\tInner:",innerIter
ecqmmf.optimize_marginals(negLogLikelihood, probsWarped, lambdaMeasureField, mu, maxInnerIter, tolerance)
mseFixed=ecqmmf.update_constant_models(fixed, probsWarped, meansFixed, variancesFixed)
mseWarped=ecqmmf.update_constant_models(warped, probsWarped, meansWarped, variancesWarped)
mse=np.max([mseFixed, mseWarped])
if(mse<tolerance):
break
#---given the intensity models and the measure field, compute the displacement
deltaField=meansWarped[None, None, :]-warped[:,:,None]
gradientField[:,:,0], gradientField[:,:,1]=sp.gradient(warped)
maxDisplacement=ecqmmf_reg.optimize_ECQMMF_displacement_field_2D(deltaField, gradientField, probsWarped, lambdaDisplacement, displacement, residuals, maxInnerIter, tolerance)
totalDisplacement+=displacement
if(maxDisplacement<tolerance):
break
plt.figure()
plt.subplot(2,2,1)
plt.imshow(fixed, cmap=plt.cm.gray)
plt.title("fixed")
plt.subplot(2,2,2)
plt.imshow(warped, cmap=plt.cm.gray)
plt.title("moving")
plt.subplot(2,2,3)
plt.imshow(probsFixed.dot(meansFixed), cmap=plt.cm.gray)
plt.title("E[fixed]")
plt.subplot(2,2,4)
plt.imshow(probsWarped.dot(meansWarped), cmap=plt.cm.gray)
plt.title("E[moving]")
return totalDisplacement
def estimateNewECQMMFMultimodalDeformationField2D(fixed,
moving,
nclasses,
lambdaMeasureField,
lambdaDisplacement,
mu,
maxOuterIter,
maxInnerIter,
tolerance,
previousDisplacement=None):
sh = fixed.shape
X0, X1 = np.mgrid[0:sh[0], 0:sh[1]]
displacement = np.empty(shape=(fixed.shape) + (2, ), dtype=np.float64)
gradientField = np.empty(shape=(fixed.shape) + (2, ), dtype=np.float64)
totalDisplacement = np.zeros(shape=(fixed.shape) + (2, ), dtype=np.float64)
gradientField = np.empty(shape=(fixed.shape) + (2, ), dtype=np.float64)
residuals = np.zeros_like(fixed)
warped = None
if (previousDisplacement != None):
totalDisplacement[...] = previousDisplacement
warped = ndimage.map_coordinates(
moving,
[X0 + totalDisplacement[..., 0], X1 + totalDisplacement[..., 1]],
prefilter=True)
else:
warped = moving
#run soft segmentation on the fixed image
meansFixed, variancesFixed = ecqmmf.initialize_constant_models(
fixed, nclasses)
meansFixed = np.array(meansFixed)
variancesFixed = np.array(variancesFixed)
segFixed, meansFixed, variancesFixed, probsFixed = ecqmmf.ecqmmf(
fixed, nclasses, lambdaMeasureField, mu, maxOuterIter, maxInnerIter,
tolerance)
meansFixed = np.array(meansFixed)
probsFixed = np.array(probsFixed)
#run soft segmentation on the warped image
meansWarped, variancesWarped = ecqmmf.initialize_constant_models(
warped, nclasses)
meansWarped = np.array(meansWarped)
variancesWarped = np.array(variancesWarped)
segWarped, meansWarped, variancesWarped, probsWarped = ecqmmf.ecqmmf(
warped, nclasses, lambdaMeasureField, mu, maxOuterIter, maxInnerIter,
tolerance)
meansWarped = np.array(meansWarped)
probsWarped = np.array(probsWarped)
#inicialize the joint models (solve assignment problem)
ecqmmf_reg.initialize_coupled_constant_models(probsFixed, probsWarped,
meansWarped)
#start optimization
outerIter = 0
negLogLikelihood = np.zeros_like(probsFixed)
while (outerIter < maxOuterIter):
outerIter += 1
print "Outer:", outerIter
if (outerIter > 1): #avoid warping twice at the first iteration
warped = ndimage.map_coordinates(moving, [
X0 + totalDisplacement[..., 0], X1 + totalDisplacement[..., 1]
],
prefilter=True)
movingMask = (moving > 0) * 1.0
warpedMask = ndimage.map_coordinates(
movingMask,
[X0 + totalDisplacement[..., 0], X1 + totalDisplacement[..., 1]],
order=0,
prefilter=False)
warpedMask = warpedMask.astype(np.int32)
#--- optimize the measure field and the intensity models ---
ecqmmf_reg.compute_registration_neg_log_likelihood_constant_models(
fixed, warped, meansFixed, meansWarped, negLogLikelihood)
#ecqmmf.initialize_normalized_likelihood(negLogLikelihood, probsWarped)
ecqmmf.initialize_maximum_likelihood(negLogLikelihood, probsWarped)
innerIter = 0
mse = 0
while (innerIter < maxInnerIter):
innerIter += 1
print "\tInner:", innerIter
ecqmmf.optimize_marginals(negLogLikelihood, probsWarped,
lambdaMeasureField, mu, maxInnerIter,
tolerance)
mseFixed = ecqmmf.update_constant_models(fixed, probsWarped,
meansFixed,
variancesFixed)
mseWarped = ecqmmf.update_constant_models(warped, probsWarped,
meansWarped,
variancesWarped)
mse = np.max([mseFixed, mseWarped])
if (mse < tolerance):
break
#---given the intensity models and the measure field, compute the displacement
deltaField = meansWarped[None, None, :] - warped[:, :, None]
gradientField[:, :, 0], gradientField[:, :, 1] = sp.gradient(warped)
maxDisplacement = ecqmmf_reg.optimize_ECQMMF_displacement_field_2D(
deltaField, gradientField, probsWarped, lambdaDisplacement,
displacement, residuals, maxInnerIter, tolerance)
totalDisplacement += displacement
if (maxDisplacement < tolerance):
break
plt.figure()
plt.subplot(2, 2, 1)
plt.imshow(fixed, cmap=plt.cm.gray)
plt.title("fixed")
plt.subplot(2, 2, 2)
plt.imshow(warped, cmap=plt.cm.gray)
plt.title("moving")
plt.subplot(2, 2, 3)
plt.imshow(probsFixed.dot(meansFixed), cmap=plt.cm.gray)
plt.title("E[fixed]")
plt.subplot(2, 2, 4)
plt.imshow(probsWarped.dot(meansWarped), cmap=plt.cm.gray)
plt.title("E[moving]")
return totalDisplacement
