paint.preprocessing.focal_spot_extractor

Classes

FocalSpot

Private initializer to ensure objects are only created using from_flux() or load().

Functions

`get_marker_coordinates`(→ tuple[torch.Tensor, Ellipsis])	Get the specific marker coordinates based on the target.
`convert_xy_to_enu`(→ torch.Tensor)	Convert (x, y) coordinates to an aimpoint in the target image based on marker positions.
`compute_center_of_intensity`(→ tuple[float, float])	Calculate the center of intensity (x, y) coordinates in a 2D flux image.
`detect_focal_spot`(→ FocalSpot)	Detect a focal spot within an image using UTIS focal spot segmentation.
`center_flux_image`(→ torch.Tensor)	Center a 2D flux image based on its intensity center of mass.

Module Contents

class paint.preprocessing.focal_spot_extractor.FocalSpot

Private initializer to ensure objects are only created using from_flux() or load().

Initializes attributes with default placeholder values.

flux: torch.Tensor

aim_point_image: tuple[float, float] = (0.0, 0.0)

aim_point: torch.Tensor

classmethod from_flux(flux: torch.Tensor, aim_point_image: tuple[float, float], aim_point: torch.Tensor) → FocalSpot

Create a FocalSpot object from provided data.

Parameters

fluxtorch.Tensor: Intensity image of the detected focal spot.
aim_point_imagetuple[float, float]: Detected center of intensity in image coordinates (height, width).
aim_pointtorch.Tensor: Detected center of intensity in global coordinates (E, N, U).

Returns

FocalSpot: Initialized FocalSpot object.

classmethod load(file_path: str | pathlib.Path) → FocalSpot

Load a FocalSpot object from disk.

Parameters

file_pathstr | Path: Base path for the focal spot files (without extensions).

Returns

FocalSpot: Loaded FocalSpot object.

save(save_path: pathlib.Path) → None

Save the focal spot object to disk.

Parameters

save_pathPath: Base path (without extension) to save the focal spot data.

paint.preprocessing.focal_spot_extractor.get_marker_coordinates(target: str | int) → tuple[torch.Tensor, Ellipsis]

Get the specific marker coordinates based on the target.

Parameters

targetstr | int: Target ID used to determine which marker coordinates should be used.

Returns

tuple[torch.Tensor, …]: Coordinates for the markers from the considered target.

paint.preprocessing.focal_spot_extractor.convert_xy_to_enu(aim_point_image: tuple[float, float], target: str | int) → torch.Tensor

Convert (x, y) coordinates to an aimpoint in the target image based on marker positions.

Parameters

aim_point_imagetuple[float, float]: Relative aim point coordinates (x, y).
targetstr | int: The target object containing marker information.

Returns

torch.Tensor: Aimpoint coordinates in the target image.

paint.preprocessing.focal_spot_extractor.compute_center_of_intensity(flux: torch.Tensor, threshold: float = 0.0) → tuple[float, float]

Calculate the center of intensity (x, y) coordinates in a 2D flux image.

Parameters

fluxtorch.Tensor: A 2D tensor representing the flux/intensity image.
thresholdfloat, optional: Minimum intensity value for pixels to be included in the calculation (Default: 0.0).

Returns

float: x coordinate of the center of intensity.
float: y coordinate of the center of intensity.

paint.preprocessing.focal_spot_extractor.detect_focal_spot(image: torch.Tensor, target: str | int, utis_model: torch.nn.Module) → FocalSpot

Detect a focal spot within an image using UTIS focal spot segmentation.

Parameters

imagetorch.Tensor: Cropped image containing the focal spot to be detected, with the shape (height, width).
targetstr | int: Target center and dimensions used for aimpoint calculation.
utis_modeltorch.nn.Module: UTIS model checkpoint used to compute the flux.

Returns

FocalSpot: Detected focal spot with flux, image coordinates, and global aimpoint.

paint.preprocessing.focal_spot_extractor.center_flux_image(flux: torch.Tensor, target_id: str | int, length: float = 6.0, n_grid: int = 256) → torch.Tensor

Center a 2D flux image based on its intensity center of mass.

Parameters

fluxtorch.Tensor: 2D tensor with shape [H, W] representing the flux distribution on the target surface.
target_idstr | int: Target identifier that determines scaling parameters.
lengthfloat: Physical edge length of the cropped flux image.
n_gridint: Resolution of the output centered image.

Returns

torch.Tensor: Centered flux image with shape [n_grid, n_grid].