|
Modules | |
| Items related to Line Of Responses (preliminary) | |
Classes | |
| class | stir::ArcCorrection |
| A class to arc-correct projection data. More... | |
| class | stir::Bin |
| A class for storing coordinates and value of a single projection bin. More... | |
| class | stir::DataSymmetriesForViewSegmentNumbers |
| A class for encoding/finding symmetries. Works only on ViewSegmentNumbers (instead of Bin). More... | |
| class | stir::DetectionPosition< coordT > |
| A class for storing coordinates of a detection. More... | |
| class | stir::DetectionPositionPair< coordT > |
| A class for storing 2 coordinates-sets of a detection, together with a timing-position index (for TOF), as suitable for PET. More... | |
| class | stir::GeometryBlocksOnCylindrical |
| A helper class to build the crystal map based on scanner info. More... | |
| class | stir::ProjData |
| The (abstract) base class for the projection data. More... | |
| class | stir::ProjDataFromStream |
| A class which reads/writes projection data from/to a (binary) stream. More... | |
| class | stir::GE::RDF_HDF5::ProjDataGEHDF5 |
| A class which reads projection data from a GE HDF5 sinogram file. More... | |
| class | stir::ProjDataInfo |
| An (abstract base) class that contains information on the projection data. More... | |
| class | stir::ProjDataInfoBlocksOnCylindricalNoArcCorr |
| Projection data info for data from a scanner with discrete dtectors organised by blocks. More... | |
| class | stir::ProjDataInfoCylindrical |
| projection data info for data corresponding to a 'cylindrical' sampling. More... | |
| class | stir::ProjDataInfoCylindricalArcCorr |
| Projection data info for arc-corrected data. More... | |
| class | stir::ProjDataInfoCylindricalNoArcCorr |
| Projection data info for data which are not arc-corrected. More... | |
| class | stir::ProjDataInfoGeneric |
| projection data info for data corresponding to 'Generic' sampling. More... | |
| class | stir::ProjDataInfoGenericNoArcCorr |
| Projection data info for data for a scanner with discrete detectors. More... | |
| class | stir::ProjDataInfoSubsetByView |
| Projection data info for data corresponding to a subset sampling by views. More... | |
| class | stir::ProjDataInMemory |
| A class which reads/writes projection data from/to memory. More... | |
| class | stir::ProjDataInterfile |
| A class which reads/writes projection data from/to a (binary) stream, but creates the corresponding Interfile header. More... | |
| class | stir::RelatedViewgrams< elemT > |
| A class for storing viewgrams which are related by symmetry. More... | |
| class | stir::Segment< elemT > |
| An (abstract base) class for storing 3d projection dataThis stores a subset of the data accessible via a ProjData object, where the SegmentIndices are fixed. More... | |
| class | stir::SegmentBySinogram< elemT > |
| A class for storing (3d) projection data with fixed SegmentIndices. More... | |
| class | stir::SegmentByView< elemT > |
| A class for storing (3d) projection data with fixed SegmentIndices. More... | |
| class | stir::SegmentIndices |
| A very simple class to store segment numbers and any other indices that define a segment. More... | |
| class | stir::Sinogram< elemT > |
| A class for 2d projection data. More... | |
| class | stir::SinogramIndices |
| A very simple class to store all dincies to get a (2D) Sinogram. More... | |
| class | stir::Viewgram< elemT > |
| A class for 2d projection data. More... | |
| class | stir::ViewgramIndices |
| A very simple class to store all dincies to get a (2D) Viewgram. More... | |
| class | stir::ProjDataInMemoryTests |
| Test class for ProjDataInMemory. More... | |
Functions | |
| Succeeded | stir::inverse_SSRB (ProjData &proj_data_4D, const ProjData &proj_data_3D) |
| Perform Inverse Single Slice Rebinning and write output to ProjData4D format. More... | |
| void | stir::multiply_crystal_factors (ProjData &proj_data, const Array< 2, float > &efficiencies, const float global_factor) |
| Construct proj-data as a multiple of crystal efficiencies (or singles) More... | |
| Succeeded | stir::scale_sinograms (ProjData &output_proj_data, const ProjData &input_proj_data, const Array< 2, float > scale_factors_per_sinogram) |
| apply a scale factor for every sinogram More... | |
| Array< 2, float > | stir::get_scale_factors_per_sinogram (const ProjData &numerator_proj_data, const ProjData &denominator_proj_data, const ProjData &weights_proj_data) |
| find scale factors between two different sinograms More... | |
| ProjDataInfo * | stir::SSRB (const ProjDataInfo &in_proj_data_info, const int num_segments_to_combine, const int num_views_to_combine=1, const int num_tangential_poss_to_trim=0, const int max_in_segment_num_to_process=-1, const int num_tof_bins_to_combine=1) |
| construct new ProjDataInfo that is appropriate for rebinned data More... | |
| void | stir::SSRB (const std::string &output_filename, const ProjData &in_projdata, const int num_segments_to_combine, const int num_views_to_combine=1, const int num_tangential_poss_to_trim=0, const bool do_normalisation=true, const int max_in_segment_num_to_process=-1, const int num_tof_bins_to_combine=1) |
| Perform Single Slice Rebinning and write output to file. More... | |
| void | stir::SSRB (ProjData &out_projdata, const ProjData &in_projdata, const bool do_normalisation=true) |
| Perform Single Slice Rebinning and write output to ProjData. More... | |
| std::ostream & | stir::operator<< (std::ostream &out, const Bin &bin) |
| Outputs a Bin to a stream.Output is of the form. More... | |
| template<class T > | |
| std::ostream & | stir::operator<< (std::ostream &out, const DetectionPosition< T > &det_pos) |
| Outputs a DetectionPosition to a stream.Output is of the form. More... | |
| template<class T > | |
| std::ostream & | stir::operator<< (std::ostream &out, const DetectionPositionPair< T > &det_pos) |
| Outputs a DetectionPosition to a stream.Output is of the form. More... | |
| Succeeded | stir::interpolate_projdata (ProjData &proj_data_out, const ProjData &proj_data_in, const BSpline::BSplineType spline_type, const bool remove_interleaving=false) |
| Perform B-Splines Interpolation. More... | |
| Succeeded | stir::interpolate_projdata (ProjData &proj_data_out, const ProjData &proj_data_in, const BasicCoordinate< 3, BSpline::BSplineType > &these_types, const bool remove_interleaving) |
Detailed Description
Basic support for projection data. This is the term generally used in STIR for data obtained by the scanner or immediate post-processing.
Function Documentation
◆ interpolate_projdata()
| Succeeded stir::interpolate_projdata | ( | ProjData & | proj_data_out, |
| const ProjData & | proj_data_in, | ||
| const BSpline::BSplineType | spline_type, | ||
| const bool | remove_interleaving = false |
||
| ) |
Perform B-Splines Interpolation.
- Parameters
-
[out] proj_data_out Its projection_data_info is used to determine output characteristics. Data will be 'put' in here using ProjData::set_sinogram(). [in] proj_data_in input data [in] spline_type determines which type of BSpline will be used [in] remove_interleaving The STIR implementation of interpolating 3D (for the moment) projdata is a generalisation that applies B-Splines Interpolation to projdata supposing that every dimension is a regular grid. For instance, for a 3D dataset, interpolating can produce a new expanded 3D dataset based on the given information (proj_data_out). This mostly is useful in the scatter sinogram expansion.
See STIR documentation about B-Spline interpolation or scatter correction.
- Todo:
- This currently only works for direct sinograms (i.e. segment 0).
- Warning
- Because of the boundary conditions in the B-spline interpolation, strange results can occur if the output sinogram has a larger range than the input sinogram.
Referenced by stir::ScatterEstimation::upsample_and_fit_scatter_estimate().
◆ inverse_SSRB()
Perform Inverse Single Slice Rebinning and write output to ProjData4D format.
- Parameters
-
[out] proj_data_4D Its projection_data_info is used to determine output characteristics (e.g. number of segments). Data will be 'put' in here using ProjData::set_sinogram(). [in] proj_data_3D input data
The STIR implementation of Inverse SSRB applies the inverse idea of SSRB. inverse_SSRB will produce oblique sinograms by finding the sinogram that has the same 'm'-coordinate (i.e. the intersection with the z-axis of the central LOR). In addition, if the output sinogram would lie 'half-way' 2 input sinograms, it will be set to the average of the 2 input sinograms.
Note that any oblique segments in proj_data_3D are currently ignored.
Input and output projectino data should have the same number of views and tangential positions.
References stir::VectorWithOffset< T >::at(), stir::ProjData::get_empty_sinogram(), stir::ProjData::get_max_axial_pos_num(), stir::ProjData::get_max_segment_num(), stir::ProjData::get_min_axial_pos_num(), stir::ProjData::get_min_segment_num(), stir::ProjData::get_proj_data_info_sptr(), and stir::warning().
Referenced by stir::ScatterEstimation::upsample_and_fit_scatter_estimate().
◆ multiply_crystal_factors()
| void stir::multiply_crystal_factors | ( | ProjData & | proj_data, |
| const Array< 2, float > & | efficiencies, | ||
| const float | global_factor | ||
| ) |
Construct proj-data as a multiple of crystal efficiencies (or singles)
- Parameters
-
[in,out] proj_data projection data to write output. This needs to be an existing object as geometry will be obtained from it. [in] efficiencies array of factors, one per crystal [in] global_factor global additional factor to use
Sets proj_data(bin) to the product global_factor times the sum of efficiencies[c1]*efficiencies[c2] (with c1,c2 the crystals in the bin).
This is useful for normalisation, but also for randoms from singles.
- Warning
- If TOF data is used, each TOF bin will be set to 1/num_tof_bins the non-TOF value. This is appropriate for RFS, but would be confusing when using for normalisation.
- , the name is a bit misleading. This function does currently not multiply the existing data with the efficiencies, but overwrites it.
References stir::error(), and stir::ProjData::get_proj_data_info_sptr().
Referenced by stir::randoms_from_singles().
◆ scale_sinograms()
| Succeeded stir::scale_sinograms | ( | ProjData & | output_proj_data, |
| const ProjData & | input_proj_data, | ||
| const Array< 2, float > | scale_factors_per_sinogram | ||
| ) |
apply a scale factor for every sinogram
- Parameters
-
[out] output_proj_data is were the new sinograms will be stored [in] input_proj_data input [in] scale_factors_per_sinogram array with the scale factors. The first index corresponds to segments, the second to axial positions.
- Returns
- indicates if writing failed or not
References stir::Bin::axial_pos_num(), stir::ProjDataInfo::get_max_axial_pos_num(), stir::ProjDataInfo::get_max_segment_num(), stir::ProjDataInfo::get_min_axial_pos_num(), stir::ProjDataInfo::get_min_segment_num(), stir::ProjData::get_proj_data_info_sptr(), stir::ProjData::get_sinogram(), stir::SegmentIndices::segment_num(), and stir::ProjData::set_sinogram().
Referenced by stir::ScatterEstimation::upsample_and_fit_scatter_estimate().
◆ get_scale_factors_per_sinogram()
| Array< 2, float > stir::get_scale_factors_per_sinogram | ( | const ProjData & | numerator_proj_data, |
| const ProjData & | denominator_proj_data, | ||
| const ProjData & | weights_proj_data | ||
| ) |
find scale factors between two different sinograms
- Parameters
-
[in] numerator_proj_data input data [in] denominator_proj_data input data [in] weights_proj_data weights to be taken into account
- Returns
- scale_factors_per_sinogram array with the scale factors. The first index corresponds to segments, the second to axial positions.
scale factors are found for every sinogram such that
Currently this function sets the scale factor or a sinogram to 1 (and calls warning()) when the denominator gets too small.
References stir::Bin::axial_pos_num(), stir::ProjDataInfo::get_max_axial_pos_num(), stir::ProjDataInfo::get_max_segment_num(), stir::ProjDataInfo::get_min_axial_pos_num(), stir::ProjDataInfo::get_min_segment_num(), stir::ProjDataInfo::get_num_tangential_poss(), stir::ProjDataInfo::get_num_views(), stir::ProjData::get_proj_data_info_sptr(), stir::ProjData::get_sinogram(), stir::VectorWithOffset< T >::resize(), stir::SegmentIndices::segment_num(), stir::Array< num_dimensions, elemT >::sum(), stir::Array< 2, elemT >::sum(), and stir::warning().
Referenced by stir::ScatterEstimation::upsample_and_fit_scatter_estimate().
◆ SSRB() [1/3]
| ProjDataInfo * stir::SSRB | ( | const ProjDataInfo & | in_proj_data_info, |
| const int | num_segments_to_combine, | ||
| const int | num_views_to_combine = 1, |
||
| const int | num_tangential_poss_to_trim = 0, |
||
| const int | max_in_segment_num_to_process = -1, |
||
| const int | num_tof_bins_to_combine = 1 |
||
| ) |
construct new ProjDataInfo that is appropriate for rebinned data
- Parameters
-
in_proj_data_info input projection data information. num_segments_to_combine how many segments will be combined into 1 output segment. num_views_to_combine how many views will be combined in the output (i.e. mashing) num_tangential_poss_to_trim can be used to throw away some bins. Half of the bins will be thrown away at each 'side' of a sinogram (see below). max_in_segment_num_to_process rebinned in_proj_data only upto this segment. Default value -1 means 'do all segments'. num_tof_bins_to_combine can be used to increase TOF mashing.
The original SSRB algorithm was developed in M.E. Daube-Witherspoon and G. Muehllehner, (1987) Treatment of axial data in three-dimensional PET, J. Nucl. Med. 28, 171-1724. It essentially ignores the obliqueness of a Line of Response and moves data to the axial position in segment 0 such that z-resolution on the axis of the scanner is preserved.
The STIR implementation of SSRB is a generalisation that applies the same idea while still allowing preserving some of the obliqueness. For instance, for a dataset with 9 segments, SSRB can produce a new dataset with only 3 segments. This essentially increases the axial compression (or span in CTI terminology), see the STIR Glossary on axial compression. In addition, SSRB can introduce extra mashing (see the STIR Glossary) of the data, i.e. add views together. Finally, it can also be used to combine TOF bins together.
Here is how to determine which bins are discarded when trimming is used. For a certain num_tangential_poss, the range is from
-(num_tangential_poss/2) to -(num_tangential_poss/2) + num_tangential_poss - 1.
The new num_tangential_poss is simply set to old_num_tangential_poss - num_tang_poss_to_trim. Note that because of this, if num_tang_poss_to_trim is negative, more (zero) bins will be added.
- Warning
- in_proj_data_info has to be (at least) of type ProjDataInfoCylindrical
- This function can only handle in_proj_data_info where all segments have identical 'num_segments_to_combine'. So it cannot handle standard GE Advance data.
- Todo:
get rid of both restrictions flagged as warnings in the documentation for this function.
rename to something much more general than
SSRB
References stir::ProjDataInfo::clone(), stir::error(), stir::ProjDataInfoCylindrical::get_axial_sampling(), stir::ProjDataInfoCylindrical::get_azimuthal_angle_offset(), stir::ProjDataInfoCylindrical::get_azimuthal_angle_sampling(), stir::ProjDataInfoCylindrical::get_m(), stir::ProjDataInfo::get_max_axial_pos_num(), stir::ProjDataInfoCylindrical::get_max_ring_difference(), stir::ProjDataInfo::get_max_segment_num(), stir::ProjDataInfo::get_min_axial_pos_num(), stir::ProjDataInfoCylindrical::get_min_ring_difference(), stir::ProjDataInfo::get_num_tangential_poss(), stir::ProjDataInfo::get_num_views(), stir::ProjDataInfo::get_tof_mash_factor(), stir::ProjDataInfoCylindrical::reduce_segment_range(), stir::round(), stir::ProjDataInfoCylindrical::set_azimuthal_angle_offset(), stir::ProjDataInfoCylindrical::set_max_axial_pos_num(), stir::ProjDataInfoCylindrical::set_max_ring_difference(), stir::ProjDataInfoCylindrical::set_min_axial_pos_num(), stir::ProjDataInfoCylindrical::set_min_ring_difference(), stir::ProjDataInfo::set_num_tangential_poss(), stir::ProjDataInfoCylindrical::set_num_views(), stir::ProjDataInfoCylindrical::set_tof_mash_factor(), and stir::warning().
Referenced by stir::FBP2DReconstruction::FBP2DReconstruction(), and stir::ScatterEstimation::make_2D_projdata_sptr().
◆ SSRB() [2/3]
| void stir::SSRB | ( | const std::string & | output_filename, |
| const ProjData & | in_projdata, | ||
| const int | num_segments_to_combine, | ||
| const int | num_views_to_combine = 1, |
||
| const int | num_tangential_poss_to_trim = 0, |
||
| const bool | do_normalisation = true, |
||
| const int | max_in_segment_num_to_process = -1, |
||
| const int | num_tof_bins_to_combine = 1 |
||
| ) |
Perform Single Slice Rebinning and write output to file.
- Parameters
-
output_filename filename to write output projection data (will be in Interfile format) in_projdata input data num_segments_to_combine how many segments will be combined into 1 output segment. max_in_segment_num_to_process rebinned in_proj_data only upto this segment. Default value -1 means 'do all segments'. do_normalisation (default true) wether to normalise the output sinograms corresponding to how many input sinograms contribute to them. num_tof_bins_to_combine defaults to 1, so TOF bins are not combined.
- See also
- SSRB(ProjData& out_projdata, const ProjData& in_projdata, const bool do_normalisation = true)
References stir::ExamData::get_exam_info_sptr(), stir::ProjData::get_proj_data_info_sptr(), and stir::SSRB().
◆ SSRB() [3/3]
| void stir::SSRB | ( | ProjData & | out_projdata, |
| const ProjData & | in_projdata, | ||
| const bool | do_normalisation = true |
||
| ) |
Perform Single Slice Rebinning and write output to ProjData.
- Parameters
-
out_projdata Output projection data. Its projection_data_info is used to determine output characteristics. Data will be 'put' in here using ProjData::set_sinogram(). in_projdata input data do_normalisation (default true) wether to normalise the output sinograms corresponding to how many (ignoring TOF) input sinograms contribute to them. Note that we do not normalise according to the number of TOF bins. This is because the projectors will take the width of the TOF bin properly into account (by integration over the TOF kernel). (In contrast, in spatial direction, projectors are outputting "normalised" data, i.e. corresponding to the line integral).
- Warning
- in_projdata has to be (at least) of type ProjDataInfoCylindrical
- See also
- SSRB(const ProjDataInfo& in_proj_data_info, const int num_segments_to_combine, const int num_views_to_combine, const int num_tang_poss_to_trim, const int max_in_segment_num_to_process, const int num_tof_bins_to_combine ) for information on the rebinning.
References stir::ProjData::get_proj_data_info_sptr().
Referenced by stir::SSRB().
◆ operator<<() [1/3]
|
inline |
Outputs a Bin to a stream.Output is of the form.
[segment_num=xx,....., value=..]
References stir::Bin::axial_pos_num(), stir::Bin::get_bin_value(), stir::SegmentIndices::segment_num(), stir::Bin::tangential_pos_num(), stir::Bin::time_frame_num(), stir::SegmentIndices::timing_pos_num(), and stir::ViewgramIndices::view_num().
◆ operator<<() [2/3]
|
inline |
Outputs a DetectionPosition to a stream.Output is of the form.
[tangential=..., axial=..., radial=...]
◆ operator<<() [3/3]
|
inline |
Outputs a DetectionPosition to a stream.Output is of the form.
[pos1=..., pos2=..., timing_pos=...]
References stir::operator>>().
