Glossary
- Tracklet
A single non-dividing cell tracked over time. In graph terms, this is the connected component of a track between divisions (daughter to next parent). Tracklets can also start or end with a non-dividing cell at the beginning and end of the captured time or if the track leaves the field of view.
- Lineage
A single cell and all of its progeny. In graph terms, a connected component including divisions.
- Node
A cell in a single timpoint. For
traccuracy, the node must have a position attribute which can either be explicitly defined as a point or derived from a segmentation mask.- Edge
A connection between two nodes in different timepoints. An edge implies that the two connected nodes represent the same cell in different timepoints.
- Skip Edges
Also known as gap closing, these are edges that connect non-consecutive frames to signify a cell being occluded or missing for some frames, before the track continues.
traccuracyallows for skip edges in both the ground truth and predicted graphs. For the purposes of standard error classification, an edgeu -> vis not considered identical to a matching edgeu -> w -> v, and this will lead to errors annotated in this region. Several metrics offer options torelax_skips_gtandrelax_skips_pred, allowing the previously described case to be considered correct. The exact handling of skip edges depends on the specific track errors and metrics in question and is detailed in the other Track Error and Metrics sections.- Division
Divisions occur when a cell divides into two daughter cells. The cell that divides is also known as the parent cell and has two outgoing edges to the two daughter cells.
- Parent cell
See division.
- Successor
For a given node in the graph, the successor is the connected node in the next timepoint. If the nodes are connected by a skip edge, the successors may be more than one timepoint away from the given node. If the given node is a parent node, there may be two successors.
- Predecessor
For a given node in the graph, the predecessor is the connected node in the previous timepoint. If the nodes are connected by a skip edge, the predecessor may be more than one timepoint away from the given node.
- True positive
An object, such as a node, edge or division, that is correctly identified or matched in the ground truth and the prediction. Abbreviated TP.
- False positive
An object, such as a node, edge or division, that is present in the predicted graph, but does not match any such object in the ground truth. Abbreviated FP.
- False negative
An object, such as a node, edge or division, that is present in the ground truth graph, but does not match any such object in the prediction. Abbreviated FN.
- Recall
\( \frac{\textrm{TP}}{\textrm{TP} + \textrm{FN}} \)
- Precision
\( \frac{\textrm{TP}}{\textrm{TP} + \textrm{FP}} \)
- F1 Score
\( \frac{2 * \textrm{Recall} * \textrm{Precision}}{\textrm{Recall} + \textrm{Precision}} \)