link - Nearest neighbor linking¶

gpu_tracking.link(to_link, search_range, memory)¶

Performs optimal (as opposed to greedy) nearest neighbor linking on the provided dataframe.

Returns:

The input dataframe with an additional column “particle”, which denotes which particles are identical throughout frames.

Return type:

pandas.DataFrame

Parameters:
  • to_link – A pandas dataframe with atleast the columns “frame”, “y” and “x”.

  • search_range – The upper cutoff distance for nearest neighbor linking.

  • memory – The number of frames that a particle is allowed to disappear before reappearing, for the purposes of linking. memory = 5 means that a particle can be gone for 5 frames, reappearing in the 6th and still be linked to the track it had built previously. If it had reappeared in the 7th, it would instead have been considered a new particle. Defaults to 0, and has no effect if search_range is not set.

gpu_tracking

Navigation

  • Application
  • Python API
    • batch - Trackpy-like tracking
    • LoG - Laplacian of the Gaussian
    • characterize_points - Characterize provided points
    • connect - Associate detections from two separate videos
    • link - Nearest neighbor linking
      • link()
    • load - Load tiff or .vsi / .ets videos
    • mean_from_file - Mean across frames from disk

Related Topics

  • Documentation overview
    • Python API
      • Previous: connect - Associate detections from two separate videos
      • Next: load - Load tiff or .vsi / .ets videos

Quick search

©2023, Andreas Malthe Faber. | Powered by Sphinx 7.0.1 & Alabaster 0.7.13 | Page source