Ridge operators

Ridge filters can be used to detect ridge-like structures, such as neurites [1], tubes [2], vessels [3], wrinkles [4] or rivers.

Different ridge filters may be suited for detecting different structures, e.g., depending on contrast or noise level.

The present class of ridge filters relies on the eigenvalues of the Hessian matrix of image intensities to detect ridge structures where the intensity changes perpendicular but not along the structure.

Note that, due to edge effects, results for Meijering and Frangi filters are cropped by 4 pixels on each edge to get a proper rendering.

References

Original image, Meijering neuriteness, Sato tubeness, Frangi vesselness, Hessian vesselness
from skimage import data
from skimage import color
from skimage.filters import meijering, sato, frangi, hessian
import matplotlib.pyplot as plt


def identity(image, **kwargs):
    """Return the original image, ignoring any kwargs."""
    return image


image = color.rgb2gray(data.retina())[300:700, 700:900]
cmap = plt.cm.gray

kwargs = {'sigmas': [1], 'mode': 'reflect'}

fig, axes = plt.subplots(2, 5)
for i, black_ridges in enumerate([1, 0]):
    for j, func in enumerate([identity, meijering, sato, frangi, hessian]):
        kwargs['black_ridges'] = black_ridges
        result = func(image, **kwargs)
        axes[i, j].imshow(result, cmap=cmap, aspect='auto')
        if i == 0:
            axes[i, j].set_title(['Original\nimage', 'Meijering\nneuriteness',
                                  'Sato\ntubeness', 'Frangi\nvesselness',
                                  'Hessian\nvesselness'][j])
        if j == 0:
            axes[i, j].set_ylabel('black_ridges = ' + str(bool(black_ridges)))
        axes[i, j].set_xticks([])
        axes[i, j].set_yticks([])

plt.tight_layout()
plt.show()

Total running time of the script: ( 0 minutes 0.456 seconds)

Gallery generated by Sphinx-Gallery