Tracks 3D¶

import numpy as np

import napari


def lissajous(t):
    a = np.random.random(size=(3,)) * 80.0 - 40.0
    b = np.random.random(size=(3,)) * 0.05
    c = np.random.random(size=(3,)) * 0.1
    return (a[i] * np.cos(b[i] * t + c[i]) for i in range(3))


def tracks_3d(num_tracks=10):
    """ create 3d+t track data """
    tracks = []

    for track_id in range(num_tracks):

        # space to store the track data and features
        track = np.zeros((200, 10), dtype=np.float32)

        # time
        timestamps = np.arange(track.shape[0])
        x, y, z = lissajous(timestamps)

        track[:, 0] = track_id
        track[:, 1] = timestamps
        track[:, 2] = 50.0 + z
        track[:, 3] = 50.0 + y
        track[:, 4] = 50.0 + x

        # calculate the speed as a feature
        gz = np.gradient(track[:, 2])
        gy = np.gradient(track[:, 3])
        gx = np.gradient(track[:, 4])

        speed = np.sqrt(gx ** 2 + gy ** 2 + gz ** 2)
        distance = np.sqrt(x ** 2 + y ** 2 + z ** 2)

        track[:, 5] = gz
        track[:, 6] = gy
        track[:, 7] = gx
        track[:, 8] = speed
        track[:, 9] = distance

        tracks.append(track)

    tracks = np.concatenate(tracks, axis=0)
    data = tracks[:, :5]  # just the coordinate data

    features = {
        'time': tracks[:, 1],
        'gradient_z': tracks[:, 5],
        'gradient_y': tracks[:, 6],
        'gradient_x': tracks[:, 7],
        'speed': tracks[:, 8],
        'distance': tracks[:, 9],
    }

    graph = {}
    return data, features, graph


tracks, features, graph = tracks_3d(num_tracks=100)
vertices = tracks[:, 1:]

viewer = napari.Viewer(ndisplay=3)
viewer.add_points(vertices, size=1, name='points', opacity=0.3)
viewer.add_tracks(tracks, features=features, name='tracks')

if __name__ == '__main__':
    napari.run()

Gallery generated by Sphinx-Gallery

To screenshot

Add multiscale image