CellCycleGAN: Spatiotemporal Microscopy Image Synthesis of Cell Populations using Statistical Shape Models and Conditional GANs

TL;DR

CellCycleGAN is a novel method that synthesizes realistic 2D+t microscopy images of cell populations by combining statistical shape models with conditional GANs, aiding in training and benchmarking segmentation algorithms.

Contribution

It introduces a new approach integrating shape models with conditional GANs for realistic spatiotemporal cell image synthesis, with instance-level control over cell cycle and fluorescence.

Findings

Generated synthetic images improve training of segmentation models.

The method enables controlled variation of cell cycle stages and fluorescence.

Synthetic data can be used for benchmarking cell tracking algorithms.

Abstract

Automatic analysis of spatio-temporal microscopy images is inevitable for state-of-the-art research in the life sciences. Recent developments in deep learning provide powerful tools for automatic analyses of such image data, but heavily depend on the amount and quality of provided training data to perform well. To this end, we developed a new method for realistic generation of synthetic 2D+t microscopy image data of fluorescently labeled cellular nuclei. The method combines spatiotemporal statistical shape models of different cell cycle stages with a conditional GAN to generate time series of cell populations and provides instance-level control of cell cycle stage and the fluorescence intensity of generated cells. We show the effect of the GAN conditioning and create a set of synthetic images that can be readily used for training and benchmarking of cell segmentation and tracking approaches.