A graph-based cell tracking algorithm with few manually tunable parameters and automated segmentation error...

Area of Science:

• Computational Biology
• Image Analysis
• Cell Biology

Background:

• Quantitative analysis of cell migration relies on automatic cell segmentation and tracking.
• Existing cell tracking algorithms often require extensive manual parameter tuning or lack automatic correction for segmentation errors.
• This limits their ease of application to new datasets and increases manual curation time.

Purpose of the Study:

• To develop a novel cell tracking algorithm with minimal manually tunable parameters and automatic segmentation error correction.
• To evaluate the algorithm's performance against established methods on diverse, simulated segmentation error datasets.
• To provide a user-friendly tool for cell migration analysis that reduces manual effort.

Main Methods:

• Development of a new tracking algorithm incorporating automatic correction for false negatives, over-segmentation, and under-segmentation errors.
• Comparison with three leading algorithms from the Cell Tracking Challenge.
• Testing on datasets with simulated segmentation degradations, including mixed error types.

Main Results:

• The proposed algorithm successfully corrects various segmentation errors, including false negatives, over- and under-segmentation.
• It demonstrates superior performance on datasets with under-segmentation or mixed segmentation errors.
• The algorithm achieved top 3 rankings in the 6th Cell Tracking Challenge without additional manual tuning.

Conclusions:

• The novel algorithm offers robust automatic segmentation error correction for cell tracking.
• It significantly reduces the need for manual curation and parameter tuning.
• This advancement facilitates more efficient and accurate quantitative insights into cell migration processes.