DeepCell

DeepCell performs deep learning-based analysis of large-scale biological images for tasks including cell nuclei identification, cell tracking, and lineage reconstruction from live-cell imaging.

Key Features:

• Scalability: Dynamically adjusts compute resources to manage and process vast quantities of biological images.
• Cost Efficiency: Demonstrated processing of 1 million 1-megapixel images in approximately 5.5 hours at a cost around US$250, with possible reductions below US$100 depending on cluster configuration.
• Cloud-Native Architecture: Operates as a cloud-native application to integrate with cloud infrastructure and support dynamic resource allocation.
• High-throughput Processing: Capable of rapid processing of millions of high-resolution images for large datasets.

Scientific Applications:

• Cell tracking and lineage reconstruction: Automated extraction of cell tracks and lineage relationships from live-cell imaging experiments.
• Cell nuclei identification and analysis: Detection and analysis of cell nuclei within large biological image datasets.
• Large-scale biological image analysis: General application to processing and extracting quantitative data from extensive imaging datasets.

Methodology:

Employs deep learning algorithms to identify and analyze cell nuclei in large biological image datasets and dynamically scales compute resources to process millions of high-resolution images.