Source code for xmipp3.protocols.protocol_preprocess.protocol_create_mask3d

# -*- coding: utf-8 -*-
# **************************************************************************
# *
# * Authors:     Carlos Oscar S. Sorzano (
#                J.M. De la Rosa Trevin (
# *
# * Unidad de  Bioinformatica of Centro Nacional de Biotecnologia , CSIC
# *
# * This program is free software; you can redistribute it and/or modify
# * it under the terms of the GNU General Public License as published by
# * the Free Software Foundation; either version 2 of the License, or
# * (at your option) any later version.
# *
# * This program is distributed in the hope that it will be useful,
# * but WITHOUT ANY WARRANTY; without even the implied warranty of
# * GNU General Public License for more details.
# *
# * You should have received a copy of the GNU General Public License
# * along with this program; if not, write to the Free Software
# * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
# * 02111-1307  USA
# *
# *  All comments concerning this program package may be sent to the
# *  e-mail address ''
# *
# **************************************************************************

from pwem import emlib
from pwem.emlib.image import ImageHandler

from pyworkflow.protocol.params import PointerParam, StringParam, PathParam

from pwem.objects import VolumeMask
from pwem.protocols import ProtCreateMask3D

from xmipp3.convert import getImageLocation
from .geometrical_mask import *





[docs]class XmippProtCreateMask3D(ProtCreateMask3D, XmippGeometricalMask3D): """ Create a 3D mask. The mask can be created with a given geometrical shape (Sphere, Box, Cylinder...) or it can be obtained from operating on a 3d volume or a previous mask. """ _label = 'create 3d mask' #--------------------------- DEFINE param functions -------------------------------------------- def _defineParams(self, form): form.addSection(label='Mask generation') form.addParam('source', EnumParam, default=SOURCE_VOLUME, choices=['Volume', 'Geometry', 'Feature File' ], label='Mask source') # For volume sources isVolume = 'source==%d' % SOURCE_VOLUME form.addParam('inputVolume', PointerParam, pointerClass="Volume", label="Input volume", condition=isVolume, help="Select the volume that will be used to create the mask") form.addParam('volumeOperation', EnumParam, default=OPERATION_THRESHOLD, choices=['Threshold', 'Segment', 'Only postprocess'], label='Operation', condition=isVolume) #TODO: add wizard form.addParam('threshold', FloatParam, default=0.0, condition='volumeOperation==%d and %s' % (OPERATION_THRESHOLD, isVolume), label='Threshold') isSegmentation = 'volumeOperation==%d and %s' % (OPERATION_SEGMENT, isVolume) form.addParam('segmentationType', EnumParam, default=SEGMENTATION_DALTON, condition=isSegmentation, label='Segmentation type', choices=['Number of voxels', 'Number of aminoacids', 'Dalton mass', 'Automatic']) form.addParam('nvoxels', IntParam, condition='%s and segmentationType==%d' % (isSegmentation, SEGMENTATION_VOXELS), label='Number of voxels') form.addParam('naminoacids', IntParam, condition='%s and segmentationType==%d' % (isSegmentation, SEGMENTATION_AMINOACIDS), label='Number of aminoacids') form.addParam('dalton', FloatParam, condition='%s and segmentationType==%d' % (isSegmentation, SEGMENTATION_DALTON), label='Mass (Da)') # For geometrical sources form.addParam('samplingRate', FloatParam, default=1, condition='source==%d or source==%d' % (SOURCE_GEOMETRY, SOURCE_FEATURE_FILE), label="Sampling Rate (Å/px)") XmippGeometricalMask3D.defineParams(self, form, isGeometry='source==%d' % SOURCE_GEOMETRY, addSize=True) # Feature File isFeatureFile = 'source==%d' % SOURCE_FEATURE_FILE form.addParam('featureFilePath', PathParam, condition=isFeatureFile, label="Feature File", help="""Create a mask using a feature file. Follows an example of feature file # XMIPP_STAR_1 * # Type of feature (sph, blo, gau, Cyl, dcy, cub, ell, con)(Required) # The operation after adding the feature to the phantom (+/=) (Required) # The feature density (Required) # The feature center (Required) # The vector for special parameters of each vector (Required) # Sphere: [radius] # Blob : [radius alpha m] Gaussian : [sigma] # Cylinder : [xradius yradius height rot tilt psi] # DCylinder : [radius height separation rot tilt psi] # Cube : [xdim ydim zdim rot tilt psi] # Ellipsoid : [xradius yradius zradius rot tilt psi] # Cone : [radius height rot tilt psi] data_block1 _dimensions3D '34 34 34' _phantomBGDensity 0. _scale 1. data_block2 loop_ _featureType _featureOperation _featureDensity _featureCenter _featureSpecificVector sph + 1 '3.03623188 0.02318841 -5.04130435' '7' """ ) # Postprocessing form.addSection(label='Postprocessing') form.addParam('doSmall', BooleanParam, default=False, label='Remove small objects', help="The input mask has to be binary") form.addParam('smallSize', IntParam, default=50, label='Minimum size',condition="doSmall", help='Connected components whose size is smaller than ' 'this number in voxels will be removed') form.addParam('doBig', BooleanParam, default=False, label='Keep largest component', help="The input mask has to be binary") form.addParam('doSymmetrize', BooleanParam, default=False, label='Symmetrize mask') form.addParam('symmetry', StringParam, default='c1', label='Symmetry group',condition="doSymmetrize", help="See \n" "for a description of the symmetry groups format. \n" "If no symmetry is present, give c1") form.addParam('doMorphological', BooleanParam, default=False, label='Apply morphological operation', help="Dilation (dilate white region). \n" "Erosion (erode white region). \n" "Closing (Dilation+Erosion, removes black spots). \n" "Opening (Erosion+Dilation, removes white spots). \n") form.addParam('morphologicalOperation', EnumParam, default=MORPHOLOGY_DILATION, condition="doMorphological", choices=['dilation', 'erosion', 'closing', 'opening'], label='Operation') form.addParam('elementSize', IntParam, default=1, condition="doMorphological", label='Structural element size', help="The larger this value, the more the effect will be noticed") form.addParam('doInvert', BooleanParam, default=False, label='Invert the mask') form.addParam('doSmooth', BooleanParam, default=False, label='Smooth borders', help="Smoothing is performed by convolving the mask with a Gaussian.") form.addParam('sigmaConvolution', FloatParam, default=2, condition="doSmooth", label='Gaussian sigma (px)', help="The larger this value, the more the effect will be noticed") #--------------------------- INSERT steps functions ------------------------ def _insertAllSteps(self): self.maskFile = self._getPath('mask.vol') if self.source == SOURCE_VOLUME: self._insertFunctionStep('createMaskFromVolumeStep') elif self.source == SOURCE_GEOMETRY: self.inputVolume.set(None) self._insertFunctionStep('createMaskFromGeometryStep') elif self.source == SOURCE_FEATURE_FILE: self._insertFunctionStep('createMaskFromFeatureFile') self._insertFunctionStep('postProcessMaskStep') self._insertFunctionStep('createOutputStep') #--------------------------- STEPS functions -------------------------------
[docs] def createMaskFromVolumeStep(self): volume = self.inputVolume.get() fnVol = getImageLocation(volume) Ts = volume.getSamplingRate() if self.volumeOperation == OPERATION_THRESHOLD: self.runJob("xmipp_transform_threshold", "-i %s -o %s --select below %f --substitute binarize" % (fnVol, self.maskFile, self.threshold.get())) elif self.volumeOperation == OPERATION_SEGMENT: args="-i %s -o %s --method " % (fnVol, self.maskFile) if self.segmentationType == SEGMENTATION_VOXELS: args += "voxel_mass %d" % (self.nvoxels.get()) elif self.segmentationType == SEGMENTATION_AMINOACIDS: args += "aa_mass %d %f" % (self.naminoacids.get(), Ts) elif self.segmentationType == SEGMENTATION_DALTON: args += "dalton_mass %d %f" % (self.dalton.get(), Ts) else: args += "otsu" self.runJob("xmipp_volume_segment", args) elif self.volumeOperation == OPERATION_POSTPROCESS: ImageHandler().convert(fnVol,self.maskFile) return [self.maskFile]
[docs] def createMaskFromGeometryStep(self): # Create empty volume file with desired dimensions size = self.size.get() emlib.createEmptyFile(self.maskFile, size, size, size) # Create the mask args = '-i %s ' % self.maskFile args += XmippGeometricalMask3D.argsForTransformMask(self,size) args += ' --create_mask %s' % self.maskFile self.runJob("xmipp_transform_mask", args) return [self.maskFile]
[docs] def createMaskFromFeatureFile(self): featFileName = self.featureFilePath.get() self.runJob("xmipp_phantom_create", "-i %s -o %s" % (featFileName, self.maskFile) ) return [self.maskFile]
[docs] def postProcessMaskStep(self): if self.doSmall: self.runJob("xmipp_transform_morphology","-i %s --binaryOperation removeSmall %d" % (self.maskFile,self.smallSize.get())) if self.doBig: self.runJob("xmipp_transform_morphology","-i %s --binaryOperation keepBiggest" % self.maskFile) if self.doSymmetrize: if self.symmetry!='c1': self.runJob("xmipp_transform_symmetrize","-i %s --sym %s --dont_wrap"%(self.maskFile,self.symmetry.get())) if self.volumeOperation==OPERATION_THRESHOLD or self.volumeOperation==OPERATION_SEGMENT: self.runJob("xmipp_transform_threshold", "-i %s --select below 0.5 --substitute binarize" % self.maskFile) if self.doMorphological: self.runJob("xmipp_transform_morphology","-i %s --binaryOperation %s --size %d" % (self.maskFile, self.getEnumText('morphologicalOperation'), self.elementSize.get())) if self.doInvert: self.runJob("xmipp_image_operate","-i %s --mult -1"%self.maskFile) self.runJob("xmipp_image_operate","-i %s --plus 1"%self.maskFile) if self.doSmooth: self.runJob("xmipp_transform_filter","-i %s --fourier real_gaussian %f"%(self.maskFile,self.sigmaConvolution.get())) self.runJob("xmipp_transform_threshold", "-i %s --select below 0 --substitute value 0" % self.maskFile)
[docs] def createOutputStep(self): volMask = VolumeMask() volMask.setFileName(self.maskFile) if self.source==SOURCE_VOLUME: volMask.setSamplingRate(self.inputVolume.get().getSamplingRate()) else: volMask.setSamplingRate(self.samplingRate.get()) self._defineOutputs(outputMask=volMask) if self.source==SOURCE_VOLUME: self._defineSourceRelation(self.inputVolume, self.outputMask)
#--------------------------- INFO functions -------------------------------------------- def _summary(self): messages = [] messages.append("*Mask creation*") if self.source==SOURCE_VOLUME: if self.volumeOperation==OPERATION_THRESHOLD: messages.append(" Thresholding %f"%self.threshold.get()) elif self.volumeOperation==OPERATION_SEGMENT: if self.segmentationType==SEGMENTATION_AUTOMATIC: messages.append(" Automatically segmented") else: m=" Segmented to a mass of " if self.segmentationType==SEGMENTATION_VOXELS: m+="%d voxels"%(int(self.nvoxels.get())) elif self.segmentationType==SEGMENTATION_AMINOACIDS: m+="%d aminoacids"%(int(self.naminoacids.get())) elif self.segmentationType==SEGMENTATION_DALTON: m+="%d daltons"%(int(self.dalton.get())) messages.append(m) elif self.source==SOURCE_GEOMETRY: size = self.size.get() messages.append(" Mask of size: %d x %d x %d"%(size,size,size)) messages += XmippGeometricalMask3D.summary(self) messages.append("*Mask processing*") if self.doSmall: messages.append(" Removing components smaller than %d" % self.smallSize.get()) if self.doBig: messages.append(" Keeping largest component") if self.doSymmetrize: messages.append(" Symmetrized %s" % self.symmetry.get()) if self.doMorphological: messages.append(" Morphological operation: %s" % self.getEnumText('morphologicalOperation')) if self.doInvert: messages.append(" Inverted") if self.doSmooth: messages.append(" Smoothed (sigma=%f)"%self.sigmaConvolution.get()) return messages def _citations(self): if (self.source == SOURCE_VOLUME and self.volumeOperation == OPERATION_SEGMENT and self.segmentationType==SEGMENTATION_AUTOMATIC): return ['Otsu1979'] def _methods(self): messages = [] if self.inputVolume.get() is None: return messages messages.append("*Mask creation*") if self.source == SOURCE_VOLUME: messages.append('We processed the volume %s.'%self.inputVolume.get().getNameId()) if self.volumeOperation == OPERATION_THRESHOLD: messages.append("We thresholded it to a gray value of %f. "%self.threshold.get()) elif self.volumeOperation == OPERATION_SEGMENT: if self.segmentationType == SEGMENTATION_AUTOMATIC: messages.append("We automatically segmented it using Otsu's method [Otsu1979]") else: m="We segmented it to a mass of " if self.segmentationType == SEGMENTATION_VOXELS: m+="%d voxels"%(int(self.nvoxels.get())) elif self.segmentationType == SEGMENTATION_AMINOACIDS: m+="%d aminoacids"%(int(self.naminoacids.get())) elif self.segmentationType == SEGMENTATION_DALTON: m+="%d daltons"%(int(self.dalton.get())) messages.append(m) elif self.source == SOURCE_GEOMETRY: size=self.size.get() messages.append("We created a mask of size: %d x %d x %d voxels. " % (size, size, size)) messages += XmippGeometricalMask3D.methods(self) if self.doSmall: messages.append("We removed components smaller than %d voxels." % (self.smallSize.get())) if self.doBig: messages.append("We kept the largest component. ") if self.doSymmetrize: messages.append("We symmetrized it as %s. "%self.symmetry.get()) if self.doMorphological: messages.append("Then, we applied a morphological operation, concisely, a %s. " % self.getEnumText('morphologicalOperation')) if self.doInvert: messages.append("We inverted the mask. ") if self.doSmooth: messages.append("And, we smoothed it (sigma=%f voxels)." % self.sigmaConvolution.get()) if self.hasAttribute('outputMask'): messages.append('We refer to the output mask as %s.' % self.outputMask.getNameId()) return messages