210-CELLECT: Efficient Contrastive Cell Tracking

The article introduces CELLECT, a novel deep learning method for contrastive embedding learning designed for large-scale, efficient cell tracking. The authors developed CELLECT to overcome the challenges of high-performance and high-efficiency tracking in massive three-dimensional (3D) time-lapse microscopy datasets by relying on sparse annotations instead of extensive manual segmentation. The system employs contrastive learning to extract robust latent embeddings of diverse cellular structures, demonstrating broad generalization across different imaging modalities and species, including applications in immunology, pathology, and neuroscience. Quantitatively, CELLECT shows significantly higher accuracy and is over 50 times faster than previous state-of-the-art algorithms like linajea in benchmark tests using C. elegans datasets, enabling robust, real-time 3D analysis of complex cellular dynamics.

References:

• Zhou, Hongyu, et al. "CELLECT: contrastive embedding learning for large-scale efficient cell tracking." Nature Methods (2025): 1-12.