visualisation

image generation functions

Functions

plot_curve(curve, chunk_size[, style, destination])

plots an insertion/deletion curve of a responsibility map

plot_3d(input, ranking, ogrid[, norm])

plots a 3d grid in matplotlib given an image <path>

_transparent_cmap(cmap[, N])

Copy colormap and set alpha values

heatmap_plot(data, resp_map, colour[, path])

spectral_plot(explanation, data, ranking, colour[, ...])

surface_plot(input, args, resp_map, target[, path])

plots a 3d surface plot

overlay_grid(img[, step_count])

remove_background(data, resp_map)

Remove the background from the responsibility map if set in the Data object

voxel_plot(args, resp_map, data[, path])

Plot a 3D voxel plot of the responsibility map using plotly.

__transpose_mask(mask, mode)

generate_colours(n, colourmap)

Generate n evenly spaced RGB colours from a matplotlib colourmap.

make_composite_mask(explanations)

Creates a composite mask from a list of masks.

apply_boundaries_to_image(image, explanations, colours)

Draws the boundaries of the explanations on the image, using the provided colours.

subplot_multi_explanations(image, explanations[, ...])

__save_multi(path, explanations_subset, data, img, ...)

save_contrastive(explanation, data, args[, path])

save_complete(explanation, data, args[, path])

save_multi_explanation(explanations, data, args[, ...])

save_image(mask, data, args[, path])

plot_image_grid(images[, ncols])

Plot a grid of images

Module Contents

visualisation.plot_curve(curve, chunk_size, style='insertion', destination=None)

plots an insertion/deletion curve of a responsibility map

visualisation.plot_3d(input, ranking, ogrid, norm=255.0)

plots a 3d grid in matplotlib given an image <path> If <path> is greyscale or RGBA, it is converted to RGB for plotting.

visualisation._transparent_cmap(cmap, N=255)

Copy colormap and set alpha values

visualisation.heatmap_plot(data, resp_map, colour, path=None)
Parameters:

data (rex_xai.input.input_data.Data)

visualisation.spectral_plot(explanation, data, ranking, colour, extra=True, path=None)
Parameters:

data (rex_xai.input.input_data.Data)

visualisation.surface_plot(input, args, resp_map, target, path=None)

plots a 3d surface plot

Parameters:

  • args (rex_xai.input.config.CausalArgs)

  • resp_map (numpy.ndarray)

  • target (rex_xai.responsibility.prediction.Prediction)

visualisation.overlay_grid(img, step_count=10)
visualisation.remove_background(data, resp_map)

Remove the background from the responsibility map if set in the Data object

Parameters:

  • data (rex_xai.input.input_data.Data)

  • resp_map (numpy.ndarray)

Return type:

numpy.ndarray

visualisation.voxel_plot(args, resp_map, data, path=None)

Plot a 3D voxel plot of the responsibility map using plotly. - Assumes the data is greyscale Produces an interactive 3D plot of the data and the responsibility map.

Parameters:

  • args (rex_xai.input.config.CausalArgs)

  • resp_map (torch.Tensor)

  • data (rex_xai.input.input_data.Data)

visualisation.__transpose_mask(mask, mode)
Parameters:
Return type:

numpy.ndarray

visualisation.generate_colours(n, colourmap)

Generate n evenly spaced RGB colours from a matplotlib colourmap.

visualisation.make_composite_mask(explanations)

Creates a composite mask from a list of masks.

visualisation.apply_boundaries_to_image(image, explanations, colours)

Draws the boundaries of the explanations on the image, using the provided colours.

visualisation.subplot_multi_explanations(image, explanations, titles=None, alpha=0.5)
visualisation.__save_multi(path, explanations_subset, data, img, colours_subset, args)
visualisation.save_contrastive(explanation, data, args, path=None)
Parameters:

args (rex_xai.input.config.CausalArgs)

visualisation.save_complete(explanation, data, args, path=None)
Parameters:

args (rex_xai.input.config.CausalArgs)

visualisation.save_multi_explanation(explanations, data, args, clause=None, path=None)
Parameters:

args (rex_xai.input.config.CausalArgs)

visualisation.save_image(mask, data, args, path=None)
Parameters:

  • mask (torch.Tensor | numpy.ndarray)

  • data (rex_xai.input.input_data.Data)

  • args (rex_xai.input.config.CausalArgs)

visualisation.plot_image_grid(images, ncols=None)

Plot a grid of images