Year: 2018
Type: preprint
Abstract: In this paper, we propose a deep learning approach for image registration by predicting deformation from image appearance. Since obtaining ground-truth deformation fields for training can be challenging, we design a fully convolutional network that is subject to dual-guidance: (1) Coarse guidance using deformation fields obtained by an existing registration method; and (2) Fine guidance using image similarity. The latter guidance helps avoid overly relying on the supervision from the training deformation fields, which could be inaccurate. For effective training, we further improve the deep convolutional network with gap filling, hierarchical loss, and multi-source strategies. Experiments on a variety of datasets show promising registration accuracy and efficiency compared with state-of-the-art methods.
Source: arXiv (Cornell University)
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Awards: Analyzing Large-Scale Neuroimaging Data in Alzheimer's Disease, Infant Brain Measurement and Super-Resolution Atlas Construction, UNC/UMN Baby Connectome Project, Assessing Large-scale Brain Connectivities in Mild Cognitive Impairment, Quantifying Brain Abnormality by Multimodality Neuroimage Analysis, Diagnosis of Alzheimer's Disease Using Dynamic High-Order Brain Networks, Longitudinal Mapping of Human Brain Development in the First Years of Life, Imaging Genomics Based Brain Disease Prediction, Constructing 4-dimensional Infant Cortical Surface Atlases, Development of Robust Brain Measurement Tools Informed by Ultrahigh Field 7T MRI
OpenAlex ID: W2787426680