Processing images using SimpleITK

See also

• SimpleITK notebooks

import numpy as np
from skimage.io import imread
import napari_simpleitk_image_processing as nsitk
import stackview

blobs = imread('https://samples.fiji.sc/blobs.png')
stackview.insight(blobs)

shape (254, 256) dtype uint8 size 63.5 kB min 8 max 248

Median filter

nsitk.median_filter(blobs, 5, 5, 0)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 24 max 248

Gaussian blur

nsitk.gaussian_blur(blobs, 10, 10, 0)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 28 max 247

Edge enhancement / edge detection

nsitk.laplacian_filter(blobs)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min -88.0 max 72.0

nsitk.laplacian_of_gaussian_filter(blobs, sigma=20)

n-sitk made image shape (254, 256) dtype float32 size 254.0 kB min -0.15148893 max 0.0996153

nsitk.sobel(blobs)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 0.0 max 519.0761023202667

nsitk.gradient_magnitude(blobs)

n-sitk made image shape (254, 256) dtype float32 size 254.0 kB min 0.0 max 66.211784

nsitk.morphological_gradient(blobs)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 0 max 240

nsitk.standard_deviation_filter(blobs)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 0 max 89

nsitk.canny_edge_detection(blobs)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 0.0 max 1.0

Denoising

nsitk.bilateral_filter(blobs, radius=10)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 34 max 238

nsitk.binominal_blur_filter(blobs, repetitions=20)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 30.42419422074899 max 247.4738098724574

nsitk.curvature_flow_denoise(blobs, number_of_iterations=150)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 19.85105575700593 max 248.34350816581673

Deconvolution

kernel = np.zeros((15, 15))
kernel[7,7] = 1
kernel = nsitk.gaussian_blur(kernel, variance_x=3, variance_y=3)
kernel

n-sitk made image shape (15, 15) dtype float64 size 1.8 kB min 0.0 max 0.05936943721673174

nsitk.richardson_lucy_deconvolution(blobs, kernel)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 0.36973499065669824 max 311.62447178679207

nsitk.wiener_deconvolution(blobs, kernel)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min -626.7738812690973 max 1074.581976763582

nsitk.tikhonov_deconvolution(blobs, kernel, regularization_constant=0.5)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min -0.05634032528120539 max 172.71681721775462

Background / foreground removal

nsitk.white_top_hat(blobs)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 0.0 max 208.0

nsitk.black_top_hat(blobs)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 0.0 max 224.0

nsitk.h_maxima(blobs, height=100)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 8 max 148

nsitk.h_minima(blobs, height=100)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 108 max 248

Other filters

rescaled = nsitk.rescale_intensity(blobs)

print(rescaled.max())

rescaled

0.9999999999999999

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 0.0 max 0.9999999999999999

nsitk.adaptive_histogram_equalization(blobs, radius_x=5, radius_y=5)

n-sitk made image shape (254, 256) dtype float64 size 508.0 kB min 20.719606939426107 max 237.734319340099

nsitk.invert_intensity(blobs)

n-sitk made image shape (254, 256) dtype uint8 size 63.5 kB min 7 max 247