Ctrl+K

Image Processing Parameters

Image Processing Parameters#

The default image processing pipeline is as follows:

  1. The dataset is converted from the original format into a PATATO HDF5 file (for example from the iThera format).

patato-import-ithera /path/to/ithera/study/folder /path/to/data/folder

For the iThera importation, pass in the study folder, which contains Scan_1, Scan_2 etc. All of these will be processed, producing Scan_1.hdf5, Scan_2.hdf5 etc in the data folder.

  1. The speed of sound is set for each scan.

patato-set-speed-of-sound /path/to/data/folder 1500

  1. The image reconstruction algorithm is applied to each scan.

patato-reconstruct /path/to/data/folder

  1. Spectral unmixing can be applied.

patato-unmix /path/to/data/folder

  1. Regions of interest can be drawn over the images for further analysis.

patato-draw-roi /path/to/data/folder

  1. Time series analysis can be applied.

patato-analyse-gc /path/to/data/folder --display True
patato-analyse-dce /path/to/data/folder --display True

  1. Further analysis can be done in Python (see examples).

>>> import patato as pat
>>> dataset = pat.PAData('/path/to/data/folder/Scan_x.hdf5')
>>> dataset.set_default_recon()
>>> dataset.get_scan_reconstructions().imshow()

Reconstruction Preset Parameters#

Reconstruction parameters can be controlled by passing in –preset. The default preset is a good starting point for most basic applications. It does a backprojection with a Hilbert transform. If a json file is passed in as a preset, you can control the reconstruction parameters. For example, you can change the band-pass filter, remove the Hilbert transform, change the pixel size, or use a Model-Based algorithm (implementation in progress).

The default reconstruction preset.#
{
    "FILTER_HIGH_PASS": 5e3, // High pass filter in Hz
    "FILTER_LOW_PASS": 7e6, // Low pass filter in Hz
    "IRF": true, // Whether to do impulse response correction
    "HILBERT_TRANSFORM": true, // Whether to do a Hilbert Transform
    "INTERPOLATE_TIME": 3, // Interpolate the time axis by this factor
    "INTERPOLATE_DETECTORS": 2, // Interpolate the detector axis by this factor
    "PREPROCESSING_ALGORITHM": "Standard Preprocessor", // Which preprocessing algorithm to use
    "RECONSTRUCTION_FIELD_OF_VIEW_X": 0.024975, // Field of view in x in metres
    "RECONSTRUCTION_FIELD_OF_VIEW_Y": 0.024975, // Field of view in y in metres
    "RECONSTRUCTION_FIELD_OF_VIEW_Z": 0., // Field of view in z in metres - ignored when RECONSTRUCTION_NZ is 1
    "RECONSTRUCTION_NX": 333, // Number of pixels in x
    "RECONSTRUCTION_NY": 333, // Number of pixels in y
    "RECONSTRUCTION_NZ": 1, // Number of pixels in z
    "RECONSTRUCTION_PARAMS": {}, // Extra parameters for the reconstruction algorithm
    "RECONSTRUCTION_ALGORITHM": "Reference Backprojection" // Which reconstruction algorithm to use
}

Unmixing Preset Parameters#

Unmixing parameters can also be controlled by passing in –preset. By default, the unmixing is done with Oxyhaemoglobin and Deoxyhaemoglobin basis spectra. To unmixing for different chromophores, pass in a json file as a preset.

The default unmixing preset.#
{
    "RESOLUTION_REDUCE": 3, // The factor by which to reduce the resolution of the reconstruction to improve SNR
    "WAVELENGTH_RANGE": [700, 900], // The wavelength range to use for unmixing
    "SPECTRA": ["Hb", "HbO2"], // The chromophores to use as basis for unmixing. Could also add "ICG".
    "SO2": true, // Whether to calculate sO2 after unmixing
    "SUFFIX": "" // What label to give the unmixing with this preset (e.g. ICG) - this allows you to make sure that
    // you use the correct unmixing in your analysis. It makes no difference to the actual algorithm.
}
Contents