binding

binding ¶

T1 mapping binding adapters for the shared fitting infrastructure.

BoundSPGRModel and BoundLookLockerModel wrap the corresponding signal models together with their fixed independent variables (flip angles / TR, inversion times), producing FittableModel instances that the shared LevenbergMarquardtFitter can use.

Both provide analytical Jacobians for efficient convergence.

BoundSPGRModel ¶

BoundSPGRModel(model, flip_angles_rad, tr, fixed=None)

Bases: BaseBoundModel

SPGR model with flip angles and TR bound.

Wraps an SPGRSignalModel so the fitter only sees free parameters [T1, M0].

PARAMETER	DESCRIPTION
`model`	SPGR signal model. TYPE: `SPGRSignalModel`
`flip_angles_rad`	Flip angles in radians. TYPE: `NDArray`
`tr`	Repetition time in milliseconds. TYPE: `float`
`fixed`	Parameters to fix at constant values during fitting. TYPE: `dict[str, float] \| None` DEFAULT: `None`

ensure_device ¶

ensure_device(xp)

Transfer flip angle arrays to the target device.

predict_array_batch ¶

predict_array_batch(free_params_batch, xp)

Predict SPGR signal for a batch of voxels.

PARAMETER	DESCRIPTION
`free_params_batch`	Free parameter values, shape `(n_free, n_voxels)`. TYPE: `NDArray`
`xp`	Array module. TYPE: `module`

RETURNS	DESCRIPTION
`NDArray`	Predicted signal, shape `(n_fa, n_voxels)`.

get_initial_guess_batch ¶

get_initial_guess_batch(observed_batch, xp)

Get initial parameter guesses from data.

Uses T1=1000 ms and M0=2*max(signal) as defaults.

PARAMETER	DESCRIPTION
`observed_batch`	Observed data, shape `(n_fa, n_voxels)`. TYPE: `NDArray`
`xp`	Array module. TYPE: `module`

RETURNS	DESCRIPTION
`NDArray`	Initial guesses, shape `(n_free, n_voxels)`.

compute_jacobian_batch ¶

compute_jacobian_batch(params_batch, _predicted, xp)

Compute analytical Jacobian for SPGR model.

PARAMETER	DESCRIPTION
`params_batch`	Free parameter values, shape `(n_free, n_voxels)`. TYPE: `NDArray`
`_predicted`	Predicted signal (unused, Jacobian computed from params). TYPE: `NDArray`
`xp`	Array module. TYPE: `module`

RETURNS	DESCRIPTION
`NDArray`	Jacobian, shape `(n_free, n_fa, n_voxels)`.

BoundLookLockerModel ¶

BoundLookLockerModel(model, ti_times, fixed=None)

Bases: BaseBoundModel

Look-Locker model with inversion times bound.

Wraps a LookLockerSignalModel so the fitter only sees free parameters [T1_star, A, B].

PARAMETER	DESCRIPTION
`model`	Look-Locker signal model. TYPE: `LookLockerSignalModel`
`ti_times`	Inversion times in milliseconds. TYPE: `NDArray`
`fixed`	Parameters to fix at constant values during fitting. TYPE: `dict[str, float] \| None` DEFAULT: `None`

ensure_device ¶

ensure_device(xp)

Transfer TI times array to the target device.

predict_array_batch ¶

predict_array_batch(free_params_batch, xp)

Predict Look-Locker signal for a batch of voxels.

S(TI) = A - B * exp(-TI / T1*)

PARAMETER	DESCRIPTION
`free_params_batch`	Free parameter values, shape `(n_free, n_voxels)`. TYPE: `NDArray`
`xp`	Array module. TYPE: `module`

RETURNS	DESCRIPTION
`NDArray`	Predicted signal, shape `(n_ti, n_voxels)`.

get_initial_guess_batch ¶

get_initial_guess_batch(observed_batch, xp)

Get initial parameter guesses from data.

Uses T1*=median(TI), A=max(S), B=max(S)+min(S).

PARAMETER	DESCRIPTION
`observed_batch`	Observed data, shape `(n_ti, n_voxels)`. TYPE: `NDArray`
`xp`	Array module. TYPE: `module`

RETURNS	DESCRIPTION
`NDArray`	Initial guesses, shape `(n_free, n_voxels)`.

compute_jacobian_batch ¶

compute_jacobian_batch(params_batch, _predicted, xp)

Compute analytical Jacobian for Look-Locker model.

PARAMETER	DESCRIPTION
`params_batch`	Free parameter values, shape `(n_free, n_voxels)`. TYPE: `NDArray`
`_predicted`	Predicted signal (unused, Jacobian computed from params). TYPE: `NDArray`
`xp`	Array module. TYPE: `module`

RETURNS	DESCRIPTION
`NDArray`	Jacobian, shape `(n_free, n_ti, n_voxels)`.