liom_toolkit.segmentation.plane_segmentation module
- liom_toolkit.segmentation.plane_segmentation.erode_mask(mask: ndarray, disk_size: int = 30) ndarray
Erode the outer edge of a mask
- Parameters:
mask (np.ndarray) – The mask to erode
disk_size (int) – The size of the disk to use for erosion
- Returns:
The eroded mask
- Return type:
np.ndarray
- liom_toolkit.segmentation.plane_segmentation.estimate_tissue_mask(img: ndarray) ndarray
Estimate the tissue mask from an image. Based on function found here: https://github.com/joe-from-mtl/sbhassisant-2d-3d-registration
- Parameters:
img (np.ndarray) – The image to estimate the mask from
- Returns:
The tissue mask
- Return type:
np.ndarray
- liom_toolkit.segmentation.plane_segmentation.frangi_filter(img: ndarray, sigma_range: tuple, black_ridges: bool = False) ndarray
Apply the Frangi filter to an image
- Parameters:
img (np.ndarray) – The image to apply the filter to
sigma_range (tuple) – The range of sigmas to use (start, stop, step)
black_ridges (bool) – Whether to detect black ridges
- Returns:
The filtered image
- Return type:
np.ndarray
- liom_toolkit.segmentation.plane_segmentation.li_threshold_image(img: ndarray) ndarray
Apply the Li thresholding algorithm to an image
- Parameters:
img (np.ndarray) – The image to apply the thresholding to
- Returns:
The thresholded image
- Return type:
np.ndarray
- liom_toolkit.segmentation.plane_segmentation.remove_small_structures(img: ndarray, mask: ndarray) ndarray
Remove small structures from a mask
- Parameters:
img (np.ndarray) – The image with which the mask was generated
mask (np.ndarray) – The mask to remove small structures from
- Returns:
The mask with small structures removed
- Return type:
np.ndarray
- liom_toolkit.segmentation.plane_segmentation.sauvola_threshold_image(img: ndarray, window_size: int = 15) ndarray
Apply the Sauvola thresholding algorithm to an image
- Parameters:
img (np.ndarray) – The image to apply the thresholding to
window_size (int) – The size of the window to use for thresholding
- Returns:
The thresholded image
- Return type:
np.ndarray
- liom_toolkit.segmentation.plane_segmentation.segment_2d_image(output_dir: str, image: ndarray, name: str, frangi_sigma_range: tuple = (2, 16, 2), frangi_black_ridges: bool = False, local_threshold: bool = False, local_threshold_size: int = 15) None
Segment 2D images. Finished files are not returned due to memory concerns, but are saved to disk.
- Parameters:
output_dir (str) – The directory to save the results to
image (np.ndarray) – The image to segment
name (str) – The name of the image
frangi_sigma_range (tuple) – The range of sigmas to use for the Frangi filter
frangi_black_ridges (bool) – Whether to detect black ridges
local_threshold (bool) – Whether to use local thresholding
local_threshold_size (int) – The size of the local thresholding window, must be odd
- liom_toolkit.segmentation.plane_segmentation.subtract_background(img: ndarray, radius: int = 70) ndarray
Subtract background from image using rolling ball algorithm
- Parameters:
img (np.ndarray) – The image to subtract the background from
radius (int) – The radius of the rolling ball
- Returns:
The background subtracted image
- Return type:
np.ndarray