sportslabkit.metrics.tracking_preprocess#

Overview#

Function#

to_mot_eval_format(gt_bbdf, pred_bbdf)

Converts tracking and ground truth data to the format(dictionary)

Functions#

sportslabkit.metrics.tracking_preprocess.to_mot_eval_format(gt_bbdf: sportslabkit.BBoxDataFrame, pred_bbdf: sportslabkit.BBoxDataFrame) dict[str, Any][source]#

Converts tracking and ground truth data to the format(dictionary) required by the MOT metrics.

Parameters:
  • gt_bbdf (BBoxDataFrame) – Bbox Dataframe for ground truth tracking data.

  • pred_bbdf (BBoxDataFrame) – Bbox Dataframe for predicted tracking data.

Returns:

Dictionary containing the data required by the MOT metrics

Return type:

dict[str, Any]

Note: data is a dict containing all of the information that metrics need to perform evaluation. It contains the following fields:

[num_timesteps, num_gt_ids, num_tracker_ids, num_gt_dets, num_tracker_dets] : integers. [gt_ids, tracker_ids]: list (for each timestep) of 1D NDArrays (for each det). [gt_dets, tracker_dets]: list (for each timestep) of lists of detection masks. [similarity_scores]: list (for each timestep) of 2D NDArrays.

reference : https://github.com/JonathonLuiten/TrackEval/blob/ec237ec3ef654548fdc1fa1e100a45b31a6d4499/trackeval/datasets/mots_challenge.py