pnp_mace.utils

Utility functions.

Functions

add_noise(clean_image, noise_std[, seed])	Add Gaussian white noise to an image.
display_image(input_image[, title, vmin, ...])	Display an image in console using matplotlib.pyplot
display_image_nrmse(input_image, reference_image)	Display an image along with the NRMSE relative to the reference image in the title.
downscale(input_image, scale_factor, resample)	Downscale the image via decimation by the given factor in each direction.
load_img(path[, convert_to_gray, ...])	Load an image given a filename or url.
nrmse(image, reference)	Calculate the NRMSE of an image with respect to a reference.
stack_init_image(init_image, num_images)	Create a list from a single image.
upscale(input_image, scale_factor, resample)	Upscale the image via replication by the given factor in each direction.

pnp_mace.utils.load_img(path, convert_to_gray=True, convert_to_float=True)

Load an image given a filename or url.

Parameters

• path – data path, may be local, or url as a string beginning with http

• convert_to_gray – True to convert color images to grayscale

• convert_to_float – True to convert to floats and divide by 255

Returns

Grayscale image (numpy ndarray)

pnp_mace.utils.display_image_nrmse(input_image, reference_image, title='', cmap='gray', fig=None, ax=None)

Display an image along with the NRMSE relative to the reference image in the title.

Parameters

• input_image – image to be displayed

• reference_image – reference image for calculating nrmse of input_image

• title – title for the plot

• cmap – colormap for image display

• fig – draw in specified figure instead of creating one

• ax – plot in specified axes instead of current axes of figure

pnp_mace.utils.display_image(input_image, title=None, vmin=0, vmax=1, cmap='gray', fig=None, ax=None)

Display an image in console using matplotlib.pyplot

Parameters

• input_image – image to be displayed

• title – title for the plot

• cmap – colormap for image display

• fig – draw in specified figure instead of creating one

• ax – plot in specified axes instead of current axes of figure

pnp_mace.utils.stack_init_image(init_image, num_images)

Create a list from a single image.

Parameters

• init_image – a single image to be copied and stacked

• num_images – number of copies to be included

Returns

A list of copies of the original image (numpy ndarrays)

pnp_mace.utils.downscale(input_image, scale_factor, resample)

Downscale the image via decimation by the given factor in each direction.

Parameters

• input_image – input image as a numpy array

• scale_factor – upscale factor

• resample – interpolation type as in PIL.Image (NEAREST = NONE = 0, LANCZOS = 1, BILINEAR = 2, BICUBIC = 3, BOX = 4, HAMMING = 5)

Returns

Downscaled image (numpy ndarray)

pnp_mace.utils.upscale(input_image, scale_factor, resample)

Upscale the image via replication by the given factor in each direction.

Parameters

• input_image – input image

• scale_factor – upscale factor

• resample – interpolation type as in PIL.Image (NEAREST = NONE = 0, LANCZOS = 1, BILINEAR = 2, BICUBIC = 3, BOX = 4, HAMMING = 5)

Returns

Upscaled image (numpy ndarray)

pnp_mace.utils.nrmse(image, reference)

Calculate the NRMSE of an image with respect to a reference.

Parameters

• image – input image to be compared with reference

• reference – reference image

Returns

Root mean square error of the difference of image and reference, divided by the root mean square of the reference

pnp_mace.utils.add_noise(clean_image, noise_std, seed=None)

Add Gaussian white noise to an image.

Parameters

• clean_image – input image

• noise_std – standard deviation of noise to be added

• seed – seed for random number generator

Returns

image with noise added, clipped to valid range of values (numpy ndarray)