def predict(self, data3d, voxelsize_mm):
data3dr = qmisc.resize_to_mm(data3d, voxelsize_mm, self.working_voxelsize_mm)
fv = self._fv(data3dr)
# print "shape predict ", fv.shape,
pred = self.cl.predict(fv)
# print "predict ", pred.shape,
return qmisc.resize_to_shape(pred.reshape(data3dr.shape), data3d.shape)
def add_train_data(self, data3d, segmentation, voxelsize_mm):
data3dr = qmisc.resize_to_mm(data3d, voxelsize_mm, self.working_voxelsize_mm)
segmentationr = qmisc.resize_to_shape(segmentation, data3dr.shape)
logger.debug(str(np.unique(segmentationr)))
logger.debug(str(data3dr.shape) + str(segmentationr.shape))
self._add_to_training_data(data3dr, self.working_voxelsize_mm, segmentationr)
def predict(self, data3d, voxelsize_mm):
data3dr = qmisc.resize_to_mm(data3d, voxelsize_mm,
self.working_voxelsize_mm)
fv = self._fv(data3dr)
# print "shape predict ", fv.shape,
pred = self.cl.predict(fv)
# print "predict ", pred.shape,
return qmisc.resize_to_shape(pred.reshape(data3dr.shape), data3d.shape)
def add_train_data(self, data3d, segmentation, voxelsize_mm):
data3dr = qmisc.resize_to_mm(data3d, voxelsize_mm,
self.working_voxelsize_mm)
segmentationr = qmisc.resize_to_shape(segmentation, data3dr.shape)
logger.debug(str(np.unique(segmentationr)))
logger.debug(str(data3dr.shape) + str(segmentationr.shape))
self._add_to_training_data(data3dr, self.working_voxelsize_mm,
segmentationr)
def scores(self, data3d, voxelsize_mm):
data3dr = qmisc.resize_to_mm(data3d, voxelsize_mm, self.working_voxelsize_mm)
fv = self._fv(data3dr)
# print "shape predict ", fv.shape,
scoreslin = self.cl.scores(fv)
scores = {}
for key in scoreslin:
scores[key] = qmisc.resize_to_shape(scoreslin[key].reshape(data3dr.shape), data3d.shape)
return scores
def scores(self, data3d, voxelsize_mm):
data3dr = qmisc.resize_to_mm(data3d, voxelsize_mm,
self.working_voxelsize_mm)
fv = self._fv(data3dr)
# print "shape predict ", fv.shape,
scoreslin = self.cl.scores(fv)
scores = {}
for key in scoreslin:
scores[key] = qmisc.resize_to_shape(
scoreslin[key].reshape(data3dr.shape), data3d.shape)
return scores
def __init__(self, data3d, voxelsize_mm):
self.voxelsize_mm = np.asarray(voxelsize_mm)
self.working_vs = np.asarray([1.5, 1.5, 1.5])
if voxelsize_mm is None:
self.data3dr = data3d
else:
self.data3dr = qmisc.resize_to_mm(data3d, voxelsize_mm, self.working_vs)
self.lungs = None
self.spine = None
self.body = None
self.orig_shape = data3d.shape
self.diaphragm_mask = None
self.angle = None
def model_score_from_sliver_data(
# output_file="~/lisa_data/liver_intensity.Model.p",
sliver_reference_dir='~/data/medical/orig/sliver07/training/',
orig_pattern="*orig*[1-9].mhd",
ref_pattern="*seg*[1-9].mhd",
modelparams={},
likelihood_ratio=0.5,
savefig=False,
savefig_fn_prefix='../graphics/bn-symmetry-',
show=False,
label='',
):
"""
:param label: text label added to all records in output table
:param sliver_reference_dir:
:param orig_pattern:
:param ref_pattern:
:param modelparams:
:param likelihood_ratio: float number between 0 and 1, scalar or list. Set the segmentation threshodl
:param savefig:
:param savefig_fn_prefix:
:param show: show images
:return:
"""
import pandas as pd
from pysegbase import pycut
import sed3
import matplotlib.pyplot as plt
import volumetry_evaluation
sliver_reference_dir = op.expanduser(sliver_reference_dir)
orig_fnames = glob.glob(sliver_reference_dir + orig_pattern)
ref_fnames = glob.glob(sliver_reference_dir + ref_pattern)
orig_fnames.sort()
ref_fnames.sort()
evaluation_all = []
for oname, rname in zip(orig_fnames, ref_fnames):
print oname
data3d_orig, metadata = io3d.datareader.read(oname)
vs_mm1 = metadata['voxelsize_mm']
data3d_seg, metadata = io3d.datareader.read(rname)
vs_mm = metadata['voxelsize_mm']
mdl = pycut.Model(modelparams=modelparams)
# m0 = mdl.mdl[2]
# len(m0.means_)
vs_mmr = [1.5, 1.5, 1.5]
data3dr = qmisc.resize_to_mm(data3d_orig, vs_mm1, vs_mmr)
lik1 = mdl.likelihood_from_image(data3dr, vs_mmr, 0)
lik2 = mdl.likelihood_from_image(data3dr, vs_mmr, 1)
if np.isscalar(likelihood_ratio):
likelihood_ratio = [likelihood_ratio]
for likelihood_ratio_i in likelihood_ratio:
if (likelihood_ratio_i <= 0) or (likelihood_ratio_i >= 1.0):
logger.error("likelihood ratio should be between 0 and 1")
seg = ((likelihood_ratio_i * lik1) > ((1.0 - likelihood_ratio_i) * lik2)).astype(np.uint8)
# seg = (lik1).astype(np.uint8)
seg_orig = qmisc.resize_to_shape(seg, data3d_orig.shape)
# seg_orig = qmisc.resize_to_shape(seg, data3d_orig.shape)
if show:
plt.figure(figsize = (15,15))
sed3.show_slices(data3d_orig , seg_orig, show=False, slice_step=20)
# likres = qmisc.resize_to_shape(lik1, data3d_orig.shape)
# sed3.show_slices(likres , seg_orig, show=False, slice_step=20)
import re
numeric_label = re.search(".*g(\d+)", oname).group(1)
# plt.figure(figsize = (5,5))
if savefig:
plt.axis('off')
# plt.imshow(symmetry_img)
filename = savefig_fn_prefix + numeric_label + '-lr-' + str(likelihood_ratio_i) + ".png"
# if np.isscalar(likelihood_ratio):
# filename = filename + ''+str
plt.savefig(filename, bbox_inches='tight')
evaluation = volumetry_evaluation.compare_volumes_sliver(seg_orig, data3d_seg, vs_mm)
evaluation['likelihood_ratio'] = likelihood_ratio_i
evaluation['numeric_label'] = numeric_label
evaluation['label'] = label
evaluation_all.append(evaluation)
# print evaluation
ev = pd.DataFrame(evaluation_all)
return ev
def model_score_from_sliver_data(
# output_file="~/lisa_data/liver_intensity.Model.p",
sliver_reference_dir='~/data/medical/orig/sliver07/training/',
orig_pattern="*orig*[1-9].mhd",
ref_pattern="*seg*[1-9].mhd",
modelparams={},
likelihood_ratio=0.5,
savefig=False,
savefig_fn_prefix='../graphics/bn-symmetry-',
show=False,
label='',
):
"""
:param label: text label added to all records in output table
:param sliver_reference_dir:
:param orig_pattern:
:param ref_pattern:
:param modelparams:
:param likelihood_ratio: float number between 0 and 1, scalar or list. Set the segmentation threshodl
:param savefig:
:param savefig_fn_prefix:
:param show: show images
:return:
"""
import pandas as pd
from pysegbase import pycut
import sed3
import matplotlib.pyplot as plt
import volumetry_evaluation
sliver_reference_dir = op.expanduser(sliver_reference_dir)
orig_fnames = glob.glob(sliver_reference_dir + orig_pattern)
ref_fnames = glob.glob(sliver_reference_dir + ref_pattern)
orig_fnames.sort()
ref_fnames.sort()
evaluation_all = []
for oname, rname in zip(orig_fnames, ref_fnames):
print oname
data3d_orig, metadata = io3d.datareader.read(oname)
vs_mm1 = metadata['voxelsize_mm']
data3d_seg, metadata = io3d.datareader.read(rname)
vs_mm = metadata['voxelsize_mm']
mdl = pycut.Model(modelparams=modelparams)
# m0 = mdl.mdl[2]
# len(m0.means_)
vs_mmr = [1.5, 1.5, 1.5]
data3dr = qmisc.resize_to_mm(data3d_orig, vs_mm1, vs_mmr)
lik1 = mdl.likelihood_from_image(data3dr, vs_mmr, 0)
lik2 = mdl.likelihood_from_image(data3dr, vs_mmr, 1)
if np.isscalar(likelihood_ratio):
likelihood_ratio = [likelihood_ratio]
for likelihood_ratio_i in likelihood_ratio:
if (likelihood_ratio_i <= 0) or (likelihood_ratio_i >= 1.0):
logger.error("likelihood ratio should be between 0 and 1")
seg = ((likelihood_ratio_i * lik1) >
((1.0 - likelihood_ratio_i) * lik2)).astype(np.uint8)
# seg = (lik1).astype(np.uint8)
seg_orig = qmisc.resize_to_shape(seg, data3d_orig.shape)
# seg_orig = qmisc.resize_to_shape(seg, data3d_orig.shape)
if show:
plt.figure(figsize=(15, 15))
sed3.show_slices(data3d_orig,
seg_orig,
show=False,
slice_step=20)
# likres = qmisc.resize_to_shape(lik1, data3d_orig.shape)
# sed3.show_slices(likres , seg_orig, show=False, slice_step=20)
import re
numeric_label = re.search(".*g(\d+)", oname).group(1)
# plt.figure(figsize = (5,5))
if savefig:
plt.axis('off')
# plt.imshow(symmetry_img)
filename = savefig_fn_prefix + numeric_label + '-lr-' + str(
likelihood_ratio_i) + ".png"
# if np.isscalar(likelihood_ratio):
# filename = filename + ''+str
plt.savefig(filename, bbox_inches='tight')
evaluation = volumetry_evaluation.compare_volumes_sliver(
seg_orig, data3d_seg, vs_mm)
evaluation['likelihood_ratio'] = likelihood_ratio_i
evaluation['numeric_label'] = numeric_label
evaluation['label'] = label
evaluation_all.append(evaluation)
# print evaluation
ev = pd.DataFrame(evaluation_all)
return ev