An objective comparison of cell-tracking algorithms.

Area of Science:

• * Life Sciences, Microscopy, and Computational Biology.

Background:

• * The Cell Tracking Challenge promotes the development and objective evaluation of cell segmentation and tracking algorithms.
• * It provides a comprehensive platform with diverse datasets and participating methods to assess current state-of-the-art in bioimage analysis.

Purpose of the Study:

• * To report on the results of three editions of the Cell Tracking Challenge.
• * To objectively evaluate and rank participating cell segmentation and tracking algorithms.
• * To analyze algorithm performance based on technical, biological, and usability measures.

Main Methods:

• * Analysis of results from 21 algorithms across 13 datasets from various microscopy modalities.
• * Application of performance measures for segmentation and tracking to rank methods.
• * Evaluation of algorithms based on biological relevance and practical usability.

Main Results:

• * No single algorithm achieved fully correct solutions across all evaluation criteria.
• * Methods incorporating learning strategies or global spatiotemporal context analysis demonstrated enhanced performance.
• * Performance varied across different segmentation and tracking scenarios.

Conclusions:

• * Current cell segmentation and tracking algorithms show significant advancements but require further improvement for complete accuracy.
• * Integrating prior information via learning or utilizing global spatiotemporal context are promising strategies for enhancing algorithm performance.
• * The Cell Tracking Challenge serves as a crucial benchmark for advancing automated cell analysis in microscopy.