ivim_pipeline

ivim_pipeline ¶

IVIM analysis pipeline.

This module provides an end-to-end IVIM-DWI analysis pipeline, integrating preprocessing and bi-exponential model fitting.

The pipeline produces parameter maps: D, D*, and f.

References

.. [1] OSIPI CAPLEX, https://osipi.github.io/OSIPI_CAPLEX/ .. [2] Le Bihan D et al. (1988). Radiology 168(2):497-505. .. [3] Dickie BR et al. MRM 2024. doi:10.1002/mrm.30101

IVIMPipelineConfig `dataclass` ¶

IVIMPipelineConfig(
    fitting_method=SEGMENTED,
    b_threshold=200.0,
    normalize_signal=True,
    output_dir=None,
    bounds=None,
    initial_guess=None,
    max_iterations=500,
    tolerance=1e-06,
    bayesian_params=None,
)

Configuration for IVIM pipeline.

ATTRIBUTE	DESCRIPTION
`fitting_method`	IVIM fitting method. TYPE: `FittingMethod`
`b_threshold`	b-value threshold for segmented fitting (s/mm²). TYPE: `float`
`normalize_signal`	Whether to normalize signal by b=0. TYPE: `bool`
`output_dir`	Output directory for results. TYPE: `Path \| None`
`bounds`	Custom parameter bounds, e.g. `{"D": (1e-4, 5e-3)}`. TYPE: `dict[str, tuple[float, float]] \| None`
`initial_guess`	Custom initial parameter estimates, e.g. `{"D": 1e-3}`. TYPE: `dict[str, float] \| None`
`max_iterations`	Maximum iterations for optimization. TYPE: `int`
`tolerance`	Convergence tolerance. TYPE: `float`
`bayesian_params`	Bayesian-specific parameters (keys: `prior_std`, `noise_std`, `compute_uncertainty`). Only used when `fitting_method` is `FittingMethod.BAYESIAN`. TYPE: `dict \| None`

IVIMPipelineResult `dataclass` ¶

IVIMPipelineResult(fit_result, config)

Result of IVIM pipeline.

ATTRIBUTE	DESCRIPTION
`fit_result`	IVIM fitting results. TYPE: `IVIMFitResult`
`config`	Pipeline configuration used. TYPE: `IVIMPipelineConfig`

IVIMPipeline ¶

IVIMPipeline(config=None)

End-to-end IVIM-DWI analysis pipeline.

This pipeline performs: 1. Signal normalization (optional) 2. IVIM bi-exponential model fitting 3. Parameter map generation (D, D*, f)

Examples:

>>> from osipy.pipeline import IVIMPipeline, IVIMPipelineConfig
>>> config = IVIMPipelineConfig(fitting_method=FittingMethod.SEGMENTED)
>>> pipeline = IVIMPipeline(config)
>>> result = pipeline.run(dwi_data, b_values)

Initialize IVIM pipeline.

PARAMETER	DESCRIPTION
`config`	Pipeline configuration. TYPE: `IVIMPipelineConfig \| None` DEFAULT: `None`

run ¶

run(dwi_data, b_values, mask=None, progress_callback=None)

Run IVIM analysis pipeline.

PARAMETER	DESCRIPTION
`dwi_data`	DWI signal data, shape (..., n_b_values). TYPE: `PerfusionDataset or NDArray`
`b_values`	b-values in s/mm². TYPE: `NDArray`
`mask`	Brain/tissue mask. TYPE: `NDArray` DEFAULT: `None`
`progress_callback`	Callback for progress updates. TYPE: `Callable` DEFAULT: `None`

RETURNS	DESCRIPTION
`IVIMPipelineResult`	Pipeline results.