Visualization

Learn how to visualize CRISPy’s results in interactive 3D plots, along with the data cube from which the ridges were identified. Use the navigation bar to skip ahead to a specific topic.

Volume Rendering in 3D

Ridge visualization are more useful in the context of the original images. CRISPy provides two options for 3D volume rendering: point cloud and isosurface volume.

Point Cloud Rendering

Note

Point cloud rendering is computationally light and recommended for typical usage.

To volume render the image cube efficiently using render_point_cloud():

from crispy.visualize import render_point_cloud
fig = render_point_cloud(cube, showfig=True, bins=15, vmin=None)

The result: an interactive 3D scatter plot with semi-transparent, color-coded points representing the volume structures in the cube image.

Isosurface Rendering

To volume render with render_volume() using layers of semi-transparent isosurfaces:

from crispy.visualize import render_volume
fig = render_volume(cube, showfig=True, surface_count=10, vmin=None)

Result: a smooth 3D rendering of the data cube visualized with isosurfaces.

Tip

Minimum value: to conserve computation strategically, use a vmin value similar to the thres value used for ridge detection when visualising with render_point_cloud() and render_volume() to render only the structures that host ridges.

Marker size: depending on the size of image, adjust the size parameter in render_point_cloud() as needed to improve the appearance of the render.

Visualizing Ridges

Pre-gridded Ridges

To plot the pre-gridded ridges with ridge_trace_3D() from saved results, first create a trace object:

from crispy.visualize import ridge_trace_3D
from crispy.grid_ridge import read_table

# read the saved results
rname = 'ridges.txt'
ridge = read_table(rname, useDict=False)
x,y,z = ridge[0], ridge[1], ridge[2]

# create a trace object for plotting
trace = ridge_trace_3D(x, y, z, size=2, color='darkred')

then plot it either over the previously rendered volume:

fig.add_trace(trace)
fig.show()

or as a standalone plot:

import plotly.graph_objects as go
go.Figure(data=[trace]).show()

Gridded Skeletons

To plot ridges that has already been gridded back to the image space (i.e., skeletons) or their pruned counterparts (i.e., spines), use either mask_trace_3D() to create a trace and plot their coordinates with a 3D scatter plot over the previously rendered structures as:

from crispy.visualize import mask_trace_3D
trace = mask_trace_3D(skeleton, size=2.5, opacity=0.9)
fig.add_trace(trace)
fig.show()

or skel_volume() to plot them as isosurface volumes using the following:

from crispy.visualize import skel_volume
fig = skel_volume(skeleton, showfig=True, opacity=0.5, fig=fig)

Saving Plots

To save the results as an interactive HTML file:

fig.write_html("ridges_3D.html")

For more, see the API Reference.