Enum MaskOp 

pub enum MaskOp {
    Threshold {
        method: MaskThresholdMethod,
        value: Option<f64>,
    },
    Bet(BetParams),
    Erode {
        iterations: usize,
    },
    Dilate {
        iterations: usize,
    },
    Close {
        radius: usize,
    },
    FillHoles {
        max_size: usize,
    },
    GaussianSmooth {
        sigma_mm: f64,
    },
    SignalErode(SignalErosionParams),
    HdBet(HdBetParams),
    Rs2Net(Rs2NetParams),
    MindGrab(MindGrabParams),
}

A single mask operation (generator or refinement)

Variants

Threshold

Fields

method: MaskThresholdMethod

value: Option<f64>

Bet(BetParams)

BET brain extraction from the magnitude (a generator), with the full BetParams — fractional intensity, smoothness, gradient threshold, iterations, subdivisions and voxel scale. BetParams::default() is FSL’s defaults on the acquisition’s own geometry.

Erode

Fields

iterations: usize

Dilate

Fields

iterations: usize

Close

Fields

radius: usize

FillHoles

Fields

max_size: usize

GaussianSmooth

Fields

sigma_mm: f64

SignalErode(SignalErosionParams)

Signal-gated erosion: peel only low-signal boundary voxels (skull-base / sinus dropout) down to a depth cap. See crate::utils::morphology::signal_gated_erosion.

Gates on the magnitude image, never on the section’s input: it divides out a receive-coil bias estimate and compares against the in-mask median, which only means anything for magnitude. Errors if no magnitude is supplied, so a phase-quality mask input can still be refined with it as long as the magnitude is available.

HdBet(HdBetParams)

HD-BET deep-learning brain extraction from the magnitude (a generator, like Bet), whatever the section’s input is. Requires the onnx feature and the hd-bet model weights; see crate::models.

Rs2Net(Rs2NetParams)

RS2-Net deep-learning rodent brain extraction from the magnitude (a generator, like HdBet). Requires the onnx feature and the rs2-net model weights; see crate::models.

MindGrab(MindGrabParams)

MindGrab deep-learning brain extraction from the magnitude (a generator, like HdBet), for any contrast. Needs ScanMetadata::affine, the onnx feature and the mindgrab model weights; see crate::models.

Implementations

impl MaskOp

pub fn dl_model_id(&self) -> Option<&'static str>

Registry id of the deep-learning model this op runs, if any.

impl MaskOp

pub fn id(&self) -> &'static str

Stable id of the operation (its parameters are not part of it).

Trait Implementations

impl Clone for MaskOp

fn clone(&self) -> MaskOp

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for MaskOp

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for MaskOp

fn eq(&self, other: &MaskOp) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for MaskOp