The Right Prior for the Right Deformation: Rethinking Continuous Deformable Image Registration
2026-08-17 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors studied different ways to align medical images by looking at how the underlying assumptions (priors) of each method affect performance. They tested methods that use smooth, local changes (B-Splines) and newer neural methods on brain and lung scans. Their results showed that methods work best when their assumptions fit the kind of movement happening in the images, like smooth or complex deformations. Overall, a technique that uses multiple scales of B-Splines (MR-D-BSCP) performed best across tasks. The study suggests choosing registration methods based on how well their deformation assumptions match the actual motion.
Deformable image registrationB-Spline transformationsImplicit priorsContinuous registrationImplicit neural representation (INR)SIRENMultiresolutionBrain MRILung CTDisplacement field
Authors
Hengjie Liu, Chushu Shen, Dan Ruan, Ke Sheng
Abstract
Deformable image registration models implicitly encode deformation priors through their parametrization and optimization. In this work, we conduct a validation study on continuous registration methods to examine how these implicit priors affect performance across different registration tasks. Classic B-Spline transformations impose locality, smoothness, and scale through their control-point structure, whereas recent INR-based methods impose different priors through neural parameterization and optimization. We compare INR-Dense (IDIR), which directly models a dense displacement field using a SIREN-based INR; INR-BSCP (SINR), which predicts B-Spline control points with an INR; D-BSCP, which directly optimizes single-scale B-Spline control points; and MR-D-BSCP, which adds a multiresolution coarse-to-fine scheme. Experiments on inter-subject brain MR registration (OASIS) and intra-subject exhale-to-inhale lung CT registration (DIR-LAB 4DCT) reveal different behavior across deformation regimes. On OASIS, where deformations are moderate but locally complex, D-BSCP matches or slightly outperforms INR-BSCP, suggesting that the B-Spline parameterization accounts for much of INR-BSCP's effectiveness. On DIR-LAB 4DCT, where respiratory motion is larger and more coherent, single-scale B-Spline methods (D-BSCP and INR-BSCP) are less suitable, while INR-Dense and MR-D-BSCP are more effective. Across both tasks, MR-D-BSCP achieves the best performance among the tested continuous parameterizations. These findings highlight that registration accuracy depends strongly on matching the induced deformation prior to the target motion pattern, and support prior-deformation matching as a practical design principle for medical image registration. Our code will be available at https://github.com/HengjieLiu/RightPriorDIR.