Learning to Upscale 3D Segmentations in Neuroimaging

2025/05/27 by Xiaoling Hu, Hu, Xiaoling, Peirong Liu +8
Biochemistry, Genetics and Molecular Biology · Computer Science · #Advanced Neural Network Applications #Cell Image Analysis Techniques #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Medical Image Segmentation Techniques

paper · pdf · doi:10.48550/arxiv.2505.21697

openalex publication_date 2025/05/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/01

Abstract

Obtaining high-resolution (HR) segmentations from coarse annotations is a pervasive challenge in computer vision. Applications include inferring pixel-level segmentations from token-level labels in vision transformers, upsampling coarse masks to full resolution, and transferring annotations from legacy low-resolution (LR) datasets to modern HR imagery. These challenges are especially acute in 3D neuroimaging, where manual labeling is costly and resolutions continually increase. We propose a scalable framework that generalizes across resolutions and domains by regressing signed distance maps, enabling smooth, boundary-aware supervision. Crucially, our model predicts one class at a time, which substantially reduces memory usage during training and inference (critical for large 3D volumes) and naturally supports generalization to unseen classes. Generalization is further improved through training on synthetic, domain-randomized data. We validate our approach on ultra-high-resolution (UHR) human brain MRI (~100 μm), where most existing methods operate at 1 mm resolution. Our framework effectively upsamples such standard-resolution segmentations to UHR detail. Results on synthetic and real data demonstrate superior scalability and generalization compared to conventional segmentation methods. Code is available at: https://github.com/HuXiaoling/Learn2Upscale.

Citations

• Learn2Synth: Learning Optimal Data Synthesis Using Hypergradients for Brain Image Segmentation
• Hierarchical Uncertainty Estimation for Learning-based Registration in Neuroimaging
• CPM: Class-conditional Prompting Machine for Audio-visual Segmentation
• FlowSDF: Flow Matching for Medical Image Segmentation Using Distance Transforms
• Registration by Regression (RbR): a framework for interpretable and flexible atlas registration
• Brain-ID: Learning Contrast-agnostic Anatomical Representations for Brain Imaging
• Towards Better Gradient Consistency for Neural Signed Distance Functions via Level Set Alignment
• Micrograph segmentations for DDEVD
• Segment Anything
• Unsupervised Inference of Signed Distance Functions from Single Sparse Point Clouds without Learning Priors
• GenSDF: Two-Stage Learning of Generalizable Signed Distance Functions
• High Quality Segmentation for Ultra High-resolution Images
• Multi-Domain Incremental Learning for Semantic Segmentation
• Recurrent Mask Refinement for Few-Shot Medical Image Segmentation
• Incremental Few-Shot Instance Segmentation
• Automated Design of Deep Learning Methods for Biomedical Image Segmentation
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
• CascadePSP: Toward Class-Agnostic and Very High-Resolution Segmentation via Global and Local Refinement
• Modeling the Background for Incremental Learning in Semantic Segmentation
• SDFDiff: Differentiable Rendering of Signed Distance Fields for 3D Shape Optimization
• Deep Distance Transform for Tubular Structure Segmentation in CT Scans
• Distance transform regression for spatially-aware deep semantic segmentation
• Models Genesis: Generic Autodidactic Models for 3D Medical Image Analysis
• DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation
• Panoptic Feature Pyramid Networks
• Rethinking Atrous Convolution for Semantic Image Segmentation
• Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World
• 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation
• V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• U-Net: Convolutional Networks for Biomedical Image Segmentation
• Learning Deconvolution Network for Semantic Segmentation
• Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs
• Multi-Atlas Segmentation of Biomedical Images: A Survey
• Fully Convolutional Networks for Semantic Segmentation

Related

• REHRSeg: Unleashing the Power of Self-Supervised Super-Resolution for Resource-Efficient 3D MRI Segmentation
• End-to-end learning of brain tissue segmentation from imperfect labeling
• CORPS: Cost-free Rigorous Pseudo-labeling based on Similarity-ranking for Brain MRI Segmentation
• Transfer Learning from Partial Annotations for Whole Brain Segmentation
• Train, Learn, Expand, Repeat