skimage2.segmentation.slic#

skimage2.segmentation.slic(image, n_segments=100, compactness=10.0, max_num_iter=10, sigma=0, spacing=None, convert2lab=None, enforce_connectivity=True, min_size_factor=0.5, max_size_factor=3, slic_zero=False, start_label=1, mask=None, *, channel_axis=-1)[source]#

Segments image using k-means clustering in Color-(x,y,z) space.

Parameters:
image(M, N[, P][, C]) ndarray

Input image. Can be 2D or 3D, and grayscale or multichannel (see channel_axis parameter). Input image must either be NaN-free or the NaN’s must be masked out.

n_segmentsint, optional

The (approximate) number of labels in the segmented output image.

compactnessfloat, optional

Balances color proximity and space proximity. Higher values give more weight to space proximity, making superpixel shapes more square/cubic. In SLICO mode, this is the initial compactness. This parameter depends strongly on image contrast and on the shapes of objects in the image. We recommend exploring possible values on a log scale, e.g., 0.01, 0.1, 1, 10, 100, before refining around a chosen value.

max_num_iterint, optional

Maximum number of iterations of k-means.

sigmafloat or array-like of floats, optional

Width of Gaussian smoothing kernel for pre-processing for each dimension of the image. The same sigma is applied to each dimension in case of a scalar value. Zero means no smoothing. Note that sigma is automatically scaled if it is scalar and if a manual voxel spacing is provided (see Notes section). If sigma is array-like, its size must match image’s number of spatial dimensions.

spacingarray-like of floats, optional

The voxel spacing along each spatial dimension. By default, slic assumes uniform spacing (same voxel resolution along each spatial dimension). This parameter controls the weights of the distances along the spatial dimensions during k-means clustering.

convert2labbool, optional

Whether the input should be converted to Lab colorspace prior to segmentation. The input image must be RGB. Highly recommended. This option defaults to True when channel_axis` is not None *and* ``image.shape[-1] == 3.

enforce_connectivitybool, optional

Whether the generated segments are connected or not

min_size_factorfloat, optional

Proportion of the minimum segment size to be removed with respect to the supposed segment size `depth*width*height/n_segments`

max_size_factorfloat, optional

Proportion of the maximum connected segment size. A value of 3 works in most of the cases.

slic_zerobool, optional

Run SLIC-zero, the zero-parameter mode of SLIC. [2]

start_labelint, optional

The labels’ index start. Should be 0 or 1.

Added in version 0.17: start_label was introduced in 0.17

maskndarray, optional

If provided, superpixels are computed only where mask is True, and seed points are homogeneously distributed over the mask using a k-means clustering strategy. Mask number of dimensions must be equal to image number of spatial dimensions.

Added in version 0.17: mask was introduced in 0.17

channel_axisint or None, optional

If None, the image is assumed to be a grayscale (single channel) image. Otherwise, this parameter indicates which axis of the array corresponds to channels.

Added in version 0.19: channel_axis was added in 0.19.

Returns:
labels2D or 3D array

Integer mask indicating segment labels.

Raises:
ValueError

If convert2lab is set to True but the last array dimension is not of length 3.

ValueError

If start_label is not 0 or 1.

ValueError

If image contains unmasked NaN values.

ValueError

If image contains unmasked infinite values.

ValueError

If image is 2D but channel_axis is -1 (the default).

Notes

  • If sigma > 0, the image is smoothed using a Gaussian kernel prior to segmentation.

  • If sigma is scalar and spacing is provided, the kernel width is divided along each dimension by the spacing. For example, if sigma=1 and spacing=[5, 1, 1], the effective sigma is [0.2, 1, 1]. This ensures sensible smoothing for anisotropic images.

  • The image is rescaled to be in [0, 1] prior to processing (masked values are ignored).

  • Images of shape (M, N, 3) are interpreted as 2D RGB images by default. To interpret them as 3D with the last dimension having length 3, use channel_axis=None.

  • start_label is introduced to handle the issue [4]. Label indexing starts at 1 by default.

References

[1]

Radhakrishna Achanta, Appu Shaji, Kevin Smith, Aurelien Lucchi, Pascal Fua, and Sabine SÞsstrunk, SLIC Superpixels Compared to State-of-the-art Superpixel Methods, TPAMI, May 2012. DOI:10.1109/TPAMI.2012.120

[3]

Irving, Benjamin. “maskSLIC: regional superpixel generation with application to local pathology characterisation in medical images.”, 2016, arXiv:1606.09518

Examples

>>> from _skimage2.segmentation import slic
>>> from _skimage2.data import astronaut
>>> img = astronaut()
>>> segments = slic(img, n_segments=100, compactness=10)

Increasing the compactness parameter yields more square regions:

>>> segments = slic(img, n_segments=100, compactness=20)
On this page
Show Source