fastER

fastER performs ultrafast, trainable segmentation of cell outlines in large-scale microscopy images to enable quantitative analysis and cell counting in biological and medical research.

Key Features:

• Speed and Efficiency: Performs segmentation orders of magnitude faster than existing tools while maintaining state-of-the-art segmentation quality.
• Robustness Across Datasets: Demonstrates robustness on challenging real and synthetic microscopy images, including cases with high cell density, pronounced cell-to-cell variability, and low signal-to-noise ratios.
• Trainable Adaptability: Adapts to specific image sets through a trainable approach that learns from user-provided labels to improve segmentation accuracy.
• Versatility: Applicable to a wide range of cell types and image acquisition methods, including brightfield time-lapse microscopy.
• Scalability: Capable of processing very large image datasets efficiently for high-throughput experiments.

Scientific Applications:

• High-throughput time-lapse quantification: Segmented and counted cells in over 200,000 brightfield images (1388 × 1040 pixels) from a six-day time-lapse experiment, identifying more than 46 million single cells in approximately two and a half hours on a standard desktop computer.

Methodology:

Employs a trainable segmentation approach that learns from user-provided labels via an interactive training process to adapt to specific datasets.