topaln

topaln performs topologically aligned matching of segmented objects across consecutive frames using integer linear programming to improve linking accuracy for cell tracking in live-cell imaging.

Key Features:

• Topological alignment: Performs topological alignments to link segmentations across successive frames and address over-segmentation and under-segmentation.
• Integer linear programming matching: Models the matching problem on a bipartite graph and computes maximum-weight solutions via integer linear programming.
• Weight computation from convex hulls: Derives weights from relative overlap scores between convex hulls of segment sets to quantify pairwise correspondence.
• Scalability and efficiency: Demonstrated ability to solve the matching task rapidly on large-scale datasets such as those from a Large Scale Digital Cell Analysis System (LSDCAS).

Scientific Applications:

• Cell tracking: Links segmented objects across frames to enable accurate cell trajectory reconstruction in live-cell videos.
• Large-scale bioimage analysis: Validated on LSDCAS-derived datasets, indicating suitability for large-scale biological imaging studies.

Methodology:

Compute topological alignments between segmentations of consecutive frames; represent segment pairs as a bipartite graph; compute relative overlap scores between convex hulls of segment sets to form weights; solve the maximum-weight matching using integer linear programming.