Submodules

Summary

cmrsim.bloch.submodules.T2starDistributionModel([...])

Submodule for Bloch simulations that implements T2 start weighting by accumulating a phase per particle.

cmrsim.bloch.submodules.ConcomitantFields(...)

Submodule for Bloch simulations that evaluates the concomitant field variation in z-direction (field variation in direction of the main magnetic field) at all particle positions contained in the trajectories according to:

cmrsim.bloch.submodules.OffResonance(gamma)

Submodule for Bloch simulations that evaluates the off-resonance phase according to:

Modules

class T2starDistributionModel(device='GPU')

Submodule for Bloch simulations that implements T2 start weighting by accumulating a phase per particle. The particle properties should be assigned randomly according to a Lorentzian distribution. The phase difference is evaluated according to:

\[\delta \phi = \omega_{T_2^\*} \delta t\]

Parameters:

device (str) – name for device placing

Methods:

__call__(omega_t2s, dt, **kwargs)

Evaluates the module equation for given inputs.

Attributes:

required_quantities

Tuple of names specifying the input quantities (datasets) to run this submodule

__call__(omega_t2s: Tensor, dt: Tensor, **kwargs) → Tensor

Evaluates the module equation for given inputs.

Parameters:

• omega_t2s (Tensor) – (#particles, ) angular frequency contribution per particle by T2star effect This should be sampled using a distribution as implemented in cmrsim.utils.particle_properties.t2star_lorentzian.

• dt (Tensor) – (, ) angular frequency contribution per particle by T2star effect

Return type:

Tensor

Returns:

(1, #particles, 1) phase increment per timestep due to T2star effects

required_quantities: Tuple[str, ...] = ('trajectories', 'omega_t2s')

Tuple of names specifying the input quantities (datasets) to run this submodule

class ConcomitantFields(gamma, b0, device='GPU')

Submodule for Bloch simulations that evaluates the concomitant field variation in z-direction (field variation in direction of the main magnetic field) at all particle positions contained in the trajectories according to:

\[r = (x, y, z)^T \ G = (G_x, G_y, G_z)^T \ \Delta \phi = (\gamma \Delta t / 2B_0) \cdot \left[ (G_x z + G_z x/2)^2 + (G_y z - G_z y/2)^2 \right]\]

Parameters:

• gamma (float) – gyromagnetic ratio in MHz/T (=1/ms/mT)

• b0 (float) – in T

• device (str) –

Methods:

__call__(gradient_wave_form, trajectories, ...)

Evaluates the module equation for given inputs.

__call__(gradient_wave_form: Tensor, trajectories: Tensor, dt: Tensor, **kwargs) → Tensor

Evaluates the module equation for given inputs.

Parameters:

• gradient_wave_form (Tensor) – (#reps, #time, 3) in mT/m

• trajectories (Tensor) – (Optional[#reps/1], #batch, #time, 3) in m

• dt (Tensor) – float -> delta time on which the gradient waveform is defined (usually ms)

Return type:

Tensor

Returns:

phase increment per time-step (#reps, #batch, #time)

class OffResonance(gamma, device='GPU')

Submodule for Bloch simulations that evaluates the off-resonance phase according to:

\[\delta \phi = \gamma \Delta B \delta t\]

Note

This function currently does not support jit-compilation due to reshape inside conditionals. (Tensorflow 2.9.1)

Parameters:

• gamma (float) – in MHz/T (1/ms/mT)

• device (str) –

Methods:

__call__(off_res, dt, **kwargs)

Evaluates the phase increment per time-step for given off resonances.

__call__(off_res: Tensor, dt: Tensor, **kwargs) → Tensor

Evaluates the phase increment per time-step for given off resonances.

Parameters:

• off_res (Tensor) – (Optional[#reps], #particles, Optional[#steps], 1) in mT. If #reps dimension is not present, expands the first dimension to yield consistent output.

• dt (Tensor) – timestep in milliseconds

Return type:

Tensor

Returns:

(#reps/1, #particles, #steps)