Label2label: training a neural network to selectively restore cellular structures in fluorescence microscopy.

Lisa Sophie Kölln1,2,3, Omar Salem2,3, Jessica Valli4

  • 1University of Strathclyde, Department of Physics, Glasgow G4 0NG, UK.

Journal of Cell Science
|January 13, 2022
PubMed
Summary

We developed label2label (L2L), a novel method using convolutional neural networks (CNNs) to reduce background noise in immunofluorescence microscopy images. L2L enhances target structure contrast, improving visualization of protein localization and cellular function.