tracking (Feature)

[pr4deepr]

Ability to read tracking data so moving cells can be processed..

• Potentially use Trackmate to track MIP generated at end of lattice acquisition.

  • Other option could be Mastodon to perform on

• xml file can be generated containing centroids and time

• Importing tracks into python

  https://forum.image.sc/t/reading-trackmate-xml-file-in-python/59262/4

• Info on napari tracks layer:

https://napari.org/stable/howtos/layers/tracks.html

Example code for reading into napari:

Get xml file into a format: track_id,t,z,y,x (if not, z, then just id,t,y,x)
example (can optimize this)
Use pandas for saving table as csv

import pandas as pd
#read xml into variable tracks

df = pd.DataFrame.from_dict(tracks).T

#explode trackdata and create a column so each row is the original index value
df = df.explode('trackData').reset_index()

#convert trackData to columns
df[['t', 'x', 'y', 'z']] = pd.DataFrame(df['trackData'].tolist(), index=df.index)
df = df[['index', 't', 'z', 'y', 'x']]
#save as csv without index
df.to_csv('./tracking.csv', index=False)

Read tracking layer into napari (can only be done programmatically)

a = pd.read_csv('tracking.csv',index_col=None)
b = a[['index','t','y','x']]
tracks = np.array(b)
#remember to use scaling when adding tracks so it matches the image scale
viewer.add_tracks(tr, name='tracks',scale=[1,, 0.14499219, 0.14499219])  #(dz,dy,dx)