Quicksilver: Fast predictive image registration – A deep learning approach

Abstract

This paper introduces , a fast deformable image registration method. registration for image-pairs works by patch-wise prediction of a deformation model based directly on image appearance. A deep encoder-decoder network is used as the prediction model. While the prediction strategy is general, we focus on predictions for the Large Deformation Diffeomorphic Metric Mapping (LDDMM) model. Specifically, we predict the momentum-parameterization of LDDMM, which facilitates a patch-wise prediction strategy while maintaining the theoretical properties of LDDMM, such as guaranteed diffeomorphic mappings for sufficiently strong regularization. We also provide a probabilistic version of our prediction network which can be sampled during the testing time to calculate uncertainties in the predicted deformations. Finally, we introduce a new correction network which greatly increases the prediction accuracy of an already existing prediction network. We show experimental results for uni-modal atlas-to-image as well as uni-/multimodal image-to-image registrations. These experiments demonstrate that our method accurately predicts registrations obtained by numerical optimization, is very fast, achieves state-of-the-art registration results on four standard validation datasets, and can jointly learn an image similarity measure. is freely available as an open-source software.

Toward predicting the evolution of lung tumors during radiotherapy observed on a longitudinal MR imaging study via a deep learning algorithm

“…Many pilot investigations in medical imaging that utilize deep learning algorithms suffer from the restrictions of the small number of patients recruited in the study. To overcome the limitation of a small dataset, methodologies that utilize finer patches rather than whole images have been developed and tested in a variety of applications including neuroanatomy segmentation, cartilage voxel classification, image template matching, and assessing image deformations . Although patches are likely highly correlated, they contain sufficient features as a diversified group and provide enough data for training the neural networks and achieve reasonable accuracies .…”

Artificial intelligence in radiology

“…Deep learning methods could handle complex tissue deformations through more advanced non-rigid registration algorithms while providing better motion compensation for temporal image sequences. Studies have shown that deep learning leads to generally more consistent registrations and is an order of magnitude faster than more conventional methods 79 . Additionally, deep learning is multimodal in nature where a single shared representation among imaging modalities can be learned 80 .…”

Prospects of deep learning for medical imaging

“…Another study proposed a technique for correcting respiratory motion during free-breathing MRI using CNN [65]. Yang et al [66] introduced a novel registration framework based on CNN, Quicksilver, which performed fast predictive image registration.…”