• Corpus ID: 2102906

Adaptive Regularization of Ill-Posed Problems: Application to Non-rigid Image Registration

  title={Adaptive Regularization of Ill-Posed Problems: Application to Non-rigid Image Registration},
  author={Andriy Myronenko and Xubo B. Song},
We introduce an adaptive regularization approach. In contrast to conventional Tikhonov regularization, which specifies a fixed regularization operator, we estimate it simultaneously with parameters. From a Bayesian perspective we estimate the prior distribution on parameters assuming that it is close to some given model distribution. We constrain the prior distribution to be a Gauss-Markov random field (GMRF), which allows us to solve for the prior distribution analytically and provides a fast… 
Nonrigid medical image registration with locally linear reconstruction
A novel locally linear reconstruction based dissimilarity measurement (LLR) that outperforms the state-of-the-art approaches in registration accuracy and is presented as a multi-resolution nonrigid image registration algorithm.
Adaptive regularization level set evolution for medical image segmentation and bias field correction
Experimental results show the advantages of the proposed level-set based segmentation method in both segmentation and bias field correction accuracies as compared with the state-of-the-art approaches.
Incorporation of image data from a previous examination in 3D serial MR imaging
It is demonstrated that follow-up scans in longitudinal examinations can be significantly accelerated by using images from previous scans as priors for constrained reconstruction using only the highly undersampled k-space data.
Deep Convolutional Neural Network for Non-rigid Image Registration
The experimental results show that a CNN can be used for efficient non-rigid image registration and in significantly less computational time than a conventional Diffeomorphic Demons or Pyramiding approach.
Image Registration Based on Low Rank Matrix: Rank-Regularized SSD
This model compensates this distortion analytically and introduces rank-regularized SSD (RRSSD), a modified SSD based on singular values of difference image in mono-modal imaging that achieves clinically acceptable registration results, and outperforms other state-of-the-art similarity measures.
An image registration-based technique for noninvasive vascular elastography
A non-rigid image registration-based technique is proposed to measure the tissue displacement for a relatively large compression, and the proposed algorithm outperforms the conventional method and reduces the relative RMSE by 25%.
MR brain images registration with kernel learning on riemannian manifold
  • Wei Liu, Leiting Chen, X. Li
  • Computer Science
    2016 13th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)
  • 2016
Experimental results show that the proposed probabilistic kernel learning based Gaussian mixture distributions in medical images registration performs better than the existing method on MR brain images.
Atlas-based Reconstruction of 3 D Intensity Volumes of Lower Extremity from 2 D Calibrated X-ray Images
A new atlas-based 2D-3D reconstruction of 3D volumes of lower extremity from a pair of calibrated X-ray images was presented. The approach combines non-rigid 2D2D registration based 3D landmark
Fully automatic reconstruction of personalized 3D volumes of the proximal femur from 2D X-ray images
A new control point-based 2D–3D registration approach for a deformable registration of a 3D volumetric template to a limited number of 2D calibrated X-ray images and its application to personalized reconstruction of 3D volumes of the proximal femur is developed.
Atlas-Based 3D Intensity Volume Reconstruction of Musculoskeletal Structures in the Lower Extremity from 2D Calibrated X-Ray Images
An atlas-based 2D-3D intensity volume reconstruction approach is presented by combining a2D-2D non-rigid registration based 3D landmark reconstruction procedure with an adaptive regularization step to prevent penetration of the reconstructed femoral and tibial volumes.


Near-optimal regularization parameters for applications in computer vision
This work applies a method for choosing near-optimal regularization parameters by approximately minimizing the distance between the true solution and the family of regularized solutions.
A Fourier Domain Framework for Variational Image Registration
A novel theoretical framework defined in the frequency domain is proposed for approaching the multidimensional image registration problem, providing more efficient implementations of the most common registration methods than already existing approaches and providing an interesting framework to design tailor-made regularization models apart from the classical, spatial domain based schemes.
Numerical methods for volume preserving image registration
Image registration is one of today's challenging image processing problems, particularly in medical imaging. Since the problem is ill posed, one may like to add additional information about
A unified approach to fast image registration and a new curvature based registration technique
Abstract Image registration is central to many challenges in medical imaging today. It has a vast range of applications. The purpose of this note is twofold. First, we review some of the most
Isotropic Energies, Filters and Splines for Vector Field Regularization
  • P. Cathier, N. Ayache
  • Mathematics, Computer Science
    Journal of Mathematical Imaging and Vision
  • 2004
This paper proposes a class of separable isotropic filters generalizing Gaussian filtering to vector fields, which enables fast smoothing in the spatial domain and solves the problem of approximating a dense and a sparse displacement field at the same time.
A fast diffeomorphic image registration algorithm
DARTEL has been applied to intersubject registration of 471 whole brain images, and the resulting deformations were evaluated in terms of how well they encode the shape information necessary to separate male and female subjects and to predict the ages of the subjects.
Nonrigid registration using free-form deformations: application to breast MR images
The results clearly indicate that the proposed nonrigid registration algorithm is much better able to recover the motion and deformation of the breast than rigid or affine registration algorithms.
Non-rigid image registration: theory and practice.
The current state of the art of non-rigid registration is discussed to put on-going research in context and to highlight current and future clinical applications that might benefit from this technology.
Iconic feature based nonrigid registration: the PASHA algorithm
The notion of iconic feature based (IFB) algorithms, which lie between geometrical and standard intensity based algorithms, are introduced, and a new registration energy for IFB registration that generalizes some of the existing techniques is presented.
Numerical Methods for Image Registration
This paper summarizes the findings of the Human Neuroscanning Project on parametric image registration and non-parametric imageRegistration and describes the setting, methodology, and results that were obtained.