ispc-lab / NeuralPCI

✨ [CVPR 2023] NeuralPCI: Spatio-temporal Neural Field for 3D Point Cloud Multi-frame Non-linear Interpolation
https://dyfcalid.github.io/NeuralPCI
Apache License 2.0
38 stars 5 forks source link

Visualization Code #3

Closed jiangchaokang closed 8 months ago

jiangchaokang commented 8 months ago

Your paper features excellent visualization effects. Could you provide the image and video visualization codes related to the paper? Access to the visualization code would be extremely helpful to me. I am looking forward to your reply. @dyfcalid

dyfcalid commented 8 months ago

We use mayavi for visualization (also can be done by open3d). There's no special technique and you can follow your own needs. As for the video demo, we save the rendered images, convert them into gif/mp4 and then make it in Adobe Premiere. So there is no straightforward code available. If you still have specific questions about details, go ahead and contact me.

An example for mayavi:

import mayavi
def viz_mayavi(points, vals="distance", colormap='summer', azi=0, zen=0):
    x=points[:, 0]
    y=points[:, 1]
    z=points[:, 2]
    d=np.sqrt(x**2+y**2)

    if vals == "height":
        col = z
    else:
        col = d

    fig = mayavi.mlab.figure(bgcolor=(1, 1, 1), size=(600, 800))
    vis = mayavi.mlab.points3d(x, y, z,
                         col,
                         mode="sphere",
                         colormap=colormap,
                         figure=fig,
                         scale_factor=0.04,
                         )
    vis.glyph.scale_mode = 'data_scaling_off'
    vis.glyph.color_mode = 'color_by_scalar'
    mayavi.mlab.view(azi, zen, distance=3.5)
    mayavi.mlab.show()