Researchers have developed a novel framework for automated fiber bundle segmentation in macaque tracer histology, utilizing synthetic data generated from diffusion MRI (dMRI) tractography. This approach synthesizes 2D image patches to train a 2D U-Net model, composing realistic foreground textures with backgrounds from blockface photos and employing domain randomization for diversification. Experiments show that this method improves generalization across different brains and fiber bundle densities compared to training solely on real data, achieving state-of-the-art performance while requiring three times less manually annotated data. AI
IMPACT This method could accelerate neuroscience research by reducing the manual annotation burden for brain pathway analysis.
RANK_REASON The cluster contains an academic paper detailing a new methodology for synthetic data generation in a specific scientific domain.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
