Estimating volumetric motion in human thorax with parametric matching constraints

Abstract

In radiotherapy (RT), organ motion caused by breathing prevents accurate patient positioning, radiation dose, and target volume determination. Most of the motion-compensated trial techniques require collaboration of the patient and expensive equipment. Estimating the motion between two computed tomography (CT) three-dimensional scans at the extremes of the breathing cycle and including this information in the RT planning has been shyly considered, mainly because that is a tedious manual task. This paper proposes a method to compute in a fully automatic fashion the spatial correspondence between those sets of volumetric CT data. Given the large ambiguity present in this problem, the method aims to reduce gradually this uncertainty through two main phases: a similarity-parametrization data analysis and a projection-regularization phase. Results on a real study show a high accuracy in establishing the spatial correspondence between both sets. Embedding this method in RT planning tools is foreseen, after making some suggested improvements and proving the validity of the two-scan approach.

Deformable registration for image-guided radiation therapy

“…Most authors [20,96,21,25,97,98] used SSD as dissimilarity measure and neglected the lung density variations due to Sarrut. 20 breathing.…”

Multiple feature models for image matching

“…The main issue of this work is the constraint-forcing procedure on second stage, because this is the point where model restrictions must be applied. More details about the motion estimation algorithm can be found in [2,3,4]. The paper is organized as follows: section 2 presents the basic ideas to achieve a better image matching by using multiple models; section 3 presents comparative results to prove the validity of this approach and illustrate the estimation improvement; finally, section 4 contains the conclusions.…”

“…For each pixel in I(x, y) a Similarity Map (SM) ρ c (x, y) of (2N + 1)-pixel width is computed according to [2,3,4], which contains the similarity between a block of I(x, y) centered at the pixel c ≡ (c x , c y ) and a set of equal size blocks of J(x, y) in a (2N + 1) search length.…”

“…Parameters of both models, Ω 1 and Ω 1 , will be Ω 1 = {x 1 , y 1 } and Ω 2 = {x 2 , y 2 }. Although the projections over these models could be obtained in a direct manner, x 1 = x 1 and x 2 = x 2 , due to special conditions of this type o dual models, and for correct application of expression (1), the coordinates of the projected vectors will be assigned according to (2).…”

“…From this idea, in previous works [2,3,4] a method to estimate the mapping Ψ(x, y) for non-rigid motion was pre-1. Parametrization: From I(x, y) and J(x, y) to obtain a set of matching parametric models Ω(x).…”