AnyStar: Domain randomized universal star-convex 3D instance segmentation
This addresses the annotation and domain adaptation challenges for researchers and practitioners in bio-microscopy and radiology, offering a universal solution that reduces manual effort and improves efficiency, though it is incremental as it builds on existing generative and segmentation methods.
The authors tackled the problem of requiring extensive manual annotation and dataset-specific retraining for 3D instance segmentation of star-convex shapes across various imaging modalities by developing AnyStar, a domain-randomized generative model that simulates synthetic training data, enabling a single network to accurately segment nuclei and other structures in multiple datasets without retraining or finetuning.
Star-convex shapes arise across bio-microscopy and radiology in the form of nuclei, nodules, metastases, and other units. Existing instance segmentation networks for such structures train on densely labeled instances for each dataset, which requires substantial and often impractical manual annotation effort. Further, significant reengineering or finetuning is needed when presented with new datasets and imaging modalities due to changes in contrast, shape, orientation, resolution, and density. We present AnyStar, a domain-randomized generative model that simulates synthetic training data of blob-like objects with randomized appearance, environments, and imaging physics to train general-purpose star-convex instance segmentation networks. As a result, networks trained using our generative model do not require annotated images from unseen datasets. A single network trained on our synthesized data accurately 3D segments C. elegans and P. dumerilii nuclei in fluorescence microscopy, mouse cortical nuclei in micro-CT, zebrafish brain nuclei in EM, and placental cotyledons in human fetal MRI, all without any retraining, finetuning, transfer learning, or domain adaptation. Code is available at https://github.com/neel-dey/AnyStar.