Skip to content

Latest commit

 

History

History
39 lines (29 loc) · 1.44 KB

README.md

File metadata and controls

39 lines (29 loc) · 1.44 KB
Raw

Database Tables

Ultrack has three SQL tables used to store the tracking data, nodes is the main table, it stores the cell information and the trajectories by saving the cells' parent, overlaps stores the overlap constraints of the hierarchies, and links the association between nodes.

Their columns are described below:

nodes

  • t: Node time point;
  • id: Node id;
  • parent_id: Node id of parent from t - 1, added after tracking;
  • t_node_id: Node id at the time point, for internal use;
  • t_hier_id: Hierarchy id of node at the time point, for internal use;
  • z: z coordinate of cell centroid;
  • y: y coordinate of cell centroid;
  • x: x coordinate of cell centroid;
  • z_shift: estimated z displacement of cell centroid;
  • y_shift: estimated y displacement of cell centroid;
  • x_shift: estimated x displacement of cell centroid;
  • area: segment size;
  • selected: boolean indicating if the node is in solution;
  • pickle: pickled data structure of the segment;
  • annotation: segmentation curation label (UNKOWN, CORRECT, UNDERSEGMENTED & OVERSEGMENTED);
  • division: division curation label (UNKOWN, TRUE, FALSE);

overlaps

  • id: Overlap id;
  • node_id: Node id, matches to node.id;
  • ancestor_id: Ancestor of overlaps.node_id node, matches to node.id;

links

  • id: Link id;
  • source_id: Source node from t, matches to node.id;
  • target_id: Target node from t + 1, matchs to node.id;
  • weight: Link association score, the higher the better.