Multimodal non‐rigid registration methods based on local variability measures in computed tomography and magnetic resonance brain images

Abstract

This paper presents a novel non-rigid multimodal registration method that relies on three basic steps: first, an initial approximation of the deformation field is obtained by a parametric registration technique based on particle filtering; second, an intensity mapping based on local variability measures (LVM) is applied over the two images in order to overcome the multimodal restriction between them; and third, an optical flow method is used in an iterative way to find the remaining displacements of the deformation field. Hence the new methodology offers a solution for multimodal NRR by a quadratic optimisation over a convex surface, which allows independent motion of each pixel, in contrast to methods that parameterise the deformation space. To evaluate the proposed method, a set of magnetic resonance/computed tomography clinical studies (pre- and post-radiotherapy treatment) of three patients with cerebral tumour deformations of the brain structures was employed. The resulting registration was evaluated both qualitatively and quantitatively by standard indices of correspondence over anatomical structures of interest in radiotherapy (brain, tumour and cerebral ventricles). These results showed that one of the proposed LVM (entropy) offers a superior performance in estimating the non-rigid deformation field

Multimodal image registration based on the expectation–maximisation methodology

“…The GPRIV, FFD‐FR, FFD‐GD, and EMER methods were implemented in MATLAB, and NRRLVM in C++ by using the ITK libraries. Also, the configuration parameters of the studied algorithms were taken from the original references [16, 31, 33].…”

“…In the first part, we show results for the parametric registration, meanwhile in the second one, the elastic formulation is evaluated. Our proposals are compared with state‐of‐the‐art algorithms [16, 21, 30–32] in synthetic and practical scenarios.…”

“…Particularly, multimodal image registration belongs to the most challenging computer vision tasks at present, because real‐life variations affect intensity values that are difficult to model accurately [5, 9]. In multimodal registration, the state of the art can be classified into: (i) reduction to mono‐modal techniques [14], (ii) parametric intensity mapping approaches [15, 16], and (iii) information theoretic approaches [17, 18]. In this paper, we aim to identify the shortcomings of these techniques in order to formulate our proposed approach.…”

Medical image rigid registration using a novel binary feature descriptor and modified affine transform

“…In (1) and (2), if the obtained values for (x, y) are decimal, they are adjusted to the nearest integers. The centre pixel P m, n is chosen such a way that the coordinates of the all N neighbouring elements must lie within the dimension of the image I.…”

“…Accurate image registration leads to the legitimate investigation of tissues even in the presence of noise due to organ motion during diagnosis [1]. Although non-rigid registration [2][3][4] finds optimal spatial correspondences between different images, this mechanism is complex and time-consuming. These are the driving factors for the researchers to prefer rigid registration methods.…”