CyclePose -- Leveraging Cycle-Consistency for Annotation-Free Nuclei Segmentation in Fluorescence Microscopy
This addresses the challenge of annotation scarcity in biomedical imaging for researchers, though it is incremental as it builds on existing generative and segmentation techniques.
The paper tackles the problem of nuclei segmentation in fluorescence microscopy without annotated data by proposing CyclePose, a hybrid framework that integrates synthetic data generation and segmentation training using a CycleGAN with cycle consistency loss for self-supervision, achieving superior performance compared to other weakly or unsupervised methods on two public datasets.
In recent years, numerous neural network architectures specifically designed for the instance segmentation of nuclei in microscopic images have been released. These models embed nuclei-specific priors to outperform generic architectures like U-Nets; however, they require large annotated datasets, which are often not available. Generative models (GANs, diffusion models) have been used to compensate for this by synthesizing training data. These two-stage approaches are computationally expensive, as first a generative model and then a segmentation model has to be trained. We propose CyclePose, a hybrid framework integrating synthetic data generation and segmentation training. CyclePose builds on a CycleGAN architecture, which allows unpaired translation between microscopy images and segmentation masks. We embed a segmentation model into CycleGAN and leverage a cycle consistency loss for self-supervision. Without annotated data, CyclePose outperforms other weakly or unsupervised methods on two public datasets. Code is available at https://github.com/jonasutz/CyclePose