Welcome to pythologist-image-utilities’s documentation!

Readme File

pythologist-image-utilities

Functions for working with image data in python

About

This set of functions are used in the construction of pythologist-reader intermediate storage format, and in the analysis of image data .. finding neighbors .. or the generation of images.

Modules

pythologist_image_utilities.flood_fill(image, x, y, exit_criteria, max_depth=1000, recursion=0, visited=None, border_trim=1)[source]

There is a flood_fill in scikit-image 0.15.dev0, but it is not faster than this for this application. It may be good to revisit skikit’s implemention if it is optimized.

Parameters:
  • image (numpy.array) – a 2d numpy array image
  • x (int) – x starting coordinate
  • y (int) – y starting coordinate
  • exit_criteria (function) – a function for which to exit i.e. lambda x: x!=0
  • max_depth (int) – a maximum recurssion depth
  • recursion (int) – not set by user, used to keep track of recursion depth
  • visited (list) – list of (x,y) tuple representing coordinates that have been visited
  • border_trim (int) – the size of the border to avoid on the edge
Returns:

the filled image
Return type:

numpy.array
pythologist_image_utilities.image_edges(image, verbose=False)[source]

Take an image of cells where pixel intensitiy integer values represent cell ids (fully filled-in) and return just the edges

Parameters:
  • image (numpy.array) – A 2d numpy array of integers coding for cell IDs
  • verbose (bool) – If true output more details to stderr
Returns:

an output image of just edges
Return type:

numpy.array
pythologist_image_utilities.make_binary_image_array(np_array)[source]

Make a binary (one channel) image from a drawn color image

Parameters:np_array (numpy.array) –
Returns:an array that is 1 where something (anything) existed vs 0 where there was nothing
Return type:numpy.array
pythologist_image_utilities.map_image_ids(image, remove_zero=True)[source]

Convert an image into a list of coordinates and the id (coded by pixel integer value)

Parameters:
  • image (numpy.array) – A numpy 2d array with the integer values representing cell IDs
  • remove_zero (bool) – If True (default), remove all zero pixels
Returns:

A pandas dataframe with columns shaped as <x><y><id>
Return type:

pandas.DataFrame
pythologist_image_utilities.median_id_coordinates(np_array, exclude_points=None)[source]

Locate a coordinate near the center of each object in an image

Parameters:
  • np_array (numpy.array) – Take an image where pixels code for the IDs
  • exclude_points (list) – optional. a list of tuples of ‘x’,’y’ coordinates. to exclude from being possible median outputs
Returns:

DataFrame indexed by ID with a near median ‘x’, and median ‘y’ for that ID
Return type:

pandas.DataFrame
pythologist_image_utilities.read_tiff_stack(filename)[source]

Read in a tiff filestack into individual images and their metadata

Parameters:filename (str) – a path to a tiff file
Returns:a list of dictionary entries keyed by ‘raw_meta’ and ‘raw_image’ for each image in the tiff stack
Return type:list
pythologist_image_utilities.watershed_image(np_array, starting_points, valid_target_points, steps=1, border=1)[source]

A function for expanding a set of pixels in an image from starting_points and into valid_target_points.

Parameters:
  • np_array (numpy.array) – A 2d array of the image where comprised of integer values
  • starting_points (list) – a list of (x,y) tuples to begin filling out from. the values of these points
  • valid_target_points (list) – a list of (x,y) tuples of valid locations to expand into
  • steps (int) – the number of times to execute the watershed
  • border (int) – the distance to remain away from the edge of the image
Returns:

the image with the watershed executed
Return type:

numpy.array

Indices and tables