Advances in Medical Imaging and Geometry Processing

The field of medical imaging and geometry processing is rapidly advancing, with a focus on developing innovative methods for image segmentation, registration, and analysis. Recent research has led to the development of new deep learning models and algorithms that can accurately segment and register medical images, including those with complex anatomical structures. One of the key directions in this field is the use of geometric deep learning techniques, which can effectively capture the spatial relationships and structures present in medical images. These techniques have been applied to a range of tasks, including image segmentation, registration, and shape analysis. Another important area of research is the development of methods for deformable image registration, which can be used to align images of the same scene or object taken at different times or from different viewpoints. This has numerous applications in medical imaging, including the tracking of tissue growth or change over time. Notable papers in this area include TissUnet, which presents a deep learning model for segmenting extracranial tissues from brain MRI scans, and DeformCL, which introduces a new deformable centerline representation for vessel extraction in 3D medical images. Other noteworthy papers include UFM, which develops a unified flow and matching model for dense image correspondence, and CINeMA, which presents a novel framework for creating high-resolution, spatio-temporal, multimodal brain atlases. These advances have the potential to improve our understanding of complex medical phenomena and enable the development of more effective diagnostic and therapeutic strategies.

Sources

TissUnet: Improved Extracranial Tissue and Cranium Segmentation for Children through Adulthood

DeformCL: Learning Deformable Centerline Representation for Vessel Extraction in 3D Medical Image

Jamais Vu: Exposing the Generalization Gap in Supervised Semantic Correspondence

Geometric deep learning for local growth prediction on abdominal aortic aneurysm surfaces

UFM: A Simple Path towards Unified Dense Correspondence with Flow

A new approach for image segmentation based on diffeomorphic registration and gradient fields

ScaleLSD: Scalable Deep Line Segment Detection Streamlined

Power Diagram Enhanced Adaptive Isosurface Extraction from Signed Distance Fields

CINeMA: Conditional Implicit Neural Multi-Modal Atlas for a Spatio-Temporal Representation of the Perinatal Brain

Learning-based density-equalizing map

RealKeyMorph: Keypoints in Real-world Coordinates for Resolution-agnostic Image Registration

Unsupervised Deformable Image Registration with Structural Nonparametric Smoothing

Built with on top of