Geometry-based synthetic data generation with CycleGAN adaptation trains a 3D U-Net that achieves mean IPQ 0.22 on real myotube images, outperforming zero-shot models.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Data Synthesis Improves 3D Myotube Instance Segmentation
Geometry-based synthetic data generation with CycleGAN adaptation trains a 3D U-Net that achieves mean IPQ 0.22 on real myotube images, outperforming zero-shot models.