Image processing

This module contains operations commonly used in image processing.

morphocut.image.ExtractROI(image: T | Variable[T], regionprops: T | Variable[T], alpha=0, bg_color=0)[source]

Extract part of an image using a RegionProperties instance.

To be used in conjunction with FindRegions.

Parameters:

  • image (np.ndarray or Variable) – Image from which regions are to be extracted.

  • regionprops (RegionProperties or Variable) – RegionProperties instance returned by FindRegions.

  • alpha – 1=Background completely reset to bg_color; 0 = Background fully visible.

  • bg_color – Color for the background.

morphocut.image.FindRegions(mask: T | Variable[T], image: T | Variable[T] = None, min_area=None, max_area=None, padding=0, warn_empty=False)[source]

Stream Find regions in a mask and calculate properties.

For more information see skimage.measure.regionprops().

Note

This Node creates multiple objects per incoming object.

Parameters:

  • mask (np.ndarray or Variable) – Mask of a given image.

  • image (np.ndarray or Variable) – An image whose mask we have to find region with.

  • min_area (int) – Minimum area of the region. If the area of our prop/region is smaller than our min_area then it will discard it.

  • max_area (int) – Maximum area of the region. If the area of our prop/region is bigger than our max_area then it will discard it.

  • padding (int) – Size of the slices/regions of our image.

  • warn_empty (bool or str or Variable) – Warn for empty images (default false). If a String is supplied, it is used as an identifier for the image.

Example

mask = ...
regionsprops = FindRegions(mask)

# regionsprops: A skimage.measure.regionsprops object.
morphocut.image.Gray2RGB(image: ndarray | Variable[ndarray], keep_dtype=False)[source]

Create an RGB representation of a gray-level image.

Note

Uses the skimage library skimage.color.gray2rgb() to convert from Grayscale to RGB.

Parameters:

image (numpy.ndarray or Variable) – Gray-level input image.

Returns:

Variable[numpy.ndarray] –

The RGB image:

An array which is the same size as the input array, but with a channel dimension appended.

morphocut.image.ImageProperties(mask: T | Variable[T], image: T | Variable[T] = None)[source]

Calculate region properties for an image containing a single object.

For more information see skimage.measure.regionprops().

Parameters:

  • mask (np.ndarray or Variable) – Mask of a given image.

  • image (np.ndarray or Variable) – An image whose mask we have to find region with.

Example

image = ...
mask = image < 128
regionsprops = ImageProperties(mask, image)

# regionsprops: A skimage.measure.regionsprops object.
morphocut.image.ImageReader(fp: T | Variable[T])[source]

Read and open the image from a given path.

Use Python Imaging Library PIL to open the file from a given path.

Parameters:

fp (file or Variable) – A filename (string), pathlib.Path object or file object.

morphocut.image.ImageStats(image: T | Variable[T], name: str = '')[source]

Parse information from a path

morphocut.image.ImageWriter(fp: T | Variable[T], image: T | Variable[T], **kwargs)[source]

Write the image into the given directory path.

Use Python Imaging Library PIL to save the image in a given path.

Parameters:

  • fp (file or Variable) – A filename (string), pathlib.Path object or file object.

  • image (np.ndarray or Variable) – Image that is to be saved into a given directory.

  • **kwargs – Arguments for PIL.Image.Image.save().

morphocut.image.RGB2Gray(image: ndarray | Variable[ndarray], keep_dtype=False)[source]

Compute luminance of an RGB image using skimage.color.rgb2gray().

Parameters:

image (numpy.ndarray or Variable) – The image in RGB format.

Returns:

Variable[numpy.ndarray] –

The luminance image:

An array which is the same size as the input array, but with the channel dimension removed and dtype=float.

class morphocut.image.RegionProperties(*args, **kwargs)[source]

Like skimage.measure.RegionProperties but without storing the whole image.

Please refer to skimage.measure.regionprops for more information on the available region properties.

property image

Sliced binary region image which has the same size as bounding box.

morphocut.image.RescaleIntensity(image: T | Variable[T], in_range: T | Variable[T] = 'image', dtype=None)[source]

Rescale the intensities of the image.

Note

Uses the skimage library skimage.exposure.rescale_intensity().

Parameters:

  • image (np.ndarray or Variable) – An image file to be rescaled.

  • in_range ((str or 2-tuple) or Variable) – min/max as the intensity range.

  • dtype (str or Variable) – min/max of the image’s dtype as the intensity range.

Returns:

Variable[np.ndarray] – Image with intensities rescaled.

morphocut.image.ThresholdConst(image: T | Variable[T], threshold: T | Variable[T])[source]

Calculate a mask by applying a constant threshold.

The result will be image < threshold.

Parameters:

  • image (np.ndarray or Variable) – Image for which the mask is to be calculated.

  • threshold (Number or Variable) – Threshold. Image intensities less than this will be True in the result.

Returns:

Variable[np.ndarray] – Mask.