Lynnicia N. Massenburg

TileBac provides a benchmark dataset for segmenting bacterial cell envelopes in ultralow-dose cryoEM images, showing foundation models outperform convolutional neural networks.

This paper evaluates segmentation models for low- and ultralow-dose cryoEM images, finding that model choice significantly affects downstream quantitative measurements.

The paper develops a facility-scale infrastructure to standardize and analyze large-scale microscopy datasets for machine learning, enabling reproducible and scalable workflows.

This paper develops an AI-driven workflow for automated analysis of bacterial ultrastructure in cryoEM data, enabling rapid and accurate quantification of membrane thickness, flagella length, and bacterial arrangement.