Synthesis¶

Functions

void image_synthesis(Context &ctx, const NufftSynthesisOptions &opt, Dataset &dataset, std::unordered_map<std::string, std::vector<std::pair<std::size_t, const float*>>> selection, ImageProp &imageProp, const std::string &imageFileName)¶

Compute image synthesis for a dataset and multiple images based on selection.

Parameters:

ctx – [in] Context.
opt – [in] Nufft options.
dataset – [in] Dataset object.
selection – [in] Selection of pairs of sample indices and scaled eigenvalues for each images.
imageProp – [in] Image properties describing the output points.
imageFileName – [in] Image file name to write output to in hdf5 format.

struct Partition¶

#include <image_synthesis.hpp>

Public Functions

Partition() = default¶

inline Partition(Partition::Grid g)¶

inline Partition(Partition::Auto g)¶

inline Partition(Partition::None g)¶

Public Members

std::variant<Partition::Auto, Partition::None, Partition::Grid> method = Partition::Auto()¶

struct Auto¶: #include <image_synthesis.hpp>

Automatic domain partition method setting.

struct Grid¶

#include <image_synthesis.hpp>

Use regular grid domain partitioning.

Public Members

std::array<std::size_t, 3> dimensions = {1, 1, 1}¶: Grid dimension.

struct None¶: #include <image_synthesis.hpp>

Disable domain partitioning.

struct NufftSynthesisOptions¶

#include <image_synthesis.hpp>

Public Functions

inline auto set_tolerance(float tol) -> NufftSynthesisOptions&¶

Set the tolerance.

Parameters:: tol – [in] Tolerance.

inline auto set_sample_batch_size(std::optional<std::size_t> size) -> NufftSynthesisOptions&¶

Set the collection group size.

Parameters:: size – [in] Collection group size.

inline auto set_local_uvw_partition(Partition p) -> NufftSynthesisOptions&¶

Set the partitioning method for the image domain.

Parameters:: p – [in] Partition method.

inline auto set_normalize_image(bool normalize) -> NufftSynthesisOptions&¶

Set normalization of image.

Parameters:: normalize – [in] True or false.

inline auto set_apply_scaling(bool apply) -> NufftSynthesisOptions&¶

Set scaling setting.

Parameters:: apply – [in] True or false.

inline auto set_precision(BippPrecision prec) -> NufftSynthesisOptions&¶

Set floating point precision.

Parameters:: prec – [in] Precision.

Public Members

BippPrecision precision = BIPP_PRECISION_SINGLE¶: Floating point pricision to use internally.

float tolerance = 0.001f¶

The tolerance used when computing the NUFFT.

Smaller value will increase accuracy but requires more operations.

std::optional<std::size_t> sampleBatchSize = std::nullopt¶

The maximum number of collected datasets processed together.

Larger number typically improves performance but requires more memory. Internal heuristic is used if unset.

Partition localUVWPartition = Partition{Partition::Auto()}¶

The partition method used in the image domain.

Partitioning decreases memory usage, but may come with a performance penalty.

bool normalizeImage = true¶: Normalize image by the number of collect steps.

bool apply_scaling = true¶: Apply scaling defined in dataset.