Frequently Asked Questions
General Questions
What is TracksData?
TracksData is a Python library that provides a common data structure and tools for multi-object tracking. It uses a graph-based representation where objects are nodes and tracks are edges connecting objects across time.
When should I use TracksData?
TracksData is ideal for:
- Multi-object tracking in microscopy or computer vision
- Biological cell tracking and lineage analysis
- Particle tracking in physics simulations
- Any scenario requiring temporal object associations
How does TracksData differ from other tracking libraries?
TracksData focuses on providing a general unified data structure and modular components that can be combined for different tracking scenarios and scale to large datasets (e.g. millions of nodes in terabytes of 3D + time imaging data).
Which graph backend should I use?
- RustWorkXGraph: For most applications where data fits in memory
- SQLGraph: For large datasets or when you need persistent storage
- GraphView: You shouldn't instantiate this directly, it is used internally by the library when you use
graph.subgraph()
Can TracksData handle cell divisions?
Yes! The :class:tracksdata.solvers.NearestNeighborsSolver lets you defined the maximum number of children nodes and :class:tracksdata.solvers.ILPSolver specifically supports division events with configurable division costs.
How do I add custom attributes?
# Add new attribute keys to the graph
graph.add_node_attr_key("my_feature", 0.0)
graph.add_edge_attr_key("confidence", 1.0)
# Use them when adding nodes/edges
graph.add_node({"t": 0, "x": 10, "my_feature": 42.0})
graph.add_edge(source_id, target_id, {"confidence": 0.95})
How do I create custom operators?
Inherit from :class:tracksdata.edges.BaseEdgesOperator or :class:tracksdata.nodes.BaseNodesOperator and implement _add_edges_per_time or _add_nodes_per_time:
from tracksdata.nodes import BaseNodesOperator
class CustomNodes(BaseNodesOperator):
def _add_nodes_per_time(self, graph, *, t):
# Your custom logic here
pass
How do I visualize results?
TracksData provides utilities for converting to napari format: