“…So for almost every t ∈]0,T [, ψ k (t) strongly converges toψ(t) in W 1,4 (Ω, R 2 ) and det ∇ψ k (t) → k→+∞ det ∇ψ(t) in L 2 (Ω). From what was done in the stationary case [15], for almost every t ∈]0,T [,…”
Section: Theoretical Resultsmentioning
confidence: 99%
“…holds, these three elements combined allowing to handle the fidelity term. In order to deal with the nonlinearity in ∇ϕ, we propose introducing an auxiliary variable V i simulating the Jacobian deformation with a quadratic penalty method as in [15]. The decoupled problem becomes :…”
Section: Numerical Resolutionmentioning
confidence: 99%
“…We then use an alternating minimization scheme to solve the problem. We refer the reader to [15,Section 4.3. ] for an exhaustive description of the algorithm relying on the derivation of Euler-Lagrange equations solved by an L 2 gradient flow algorithm and an implicit/semi-implicit Euler time stepping scheme.…”
Section: Numerical Resolutionmentioning
confidence: 99%
“…Sharing representation between tasks and carefully intertwining them allows to reduce error propagation, to create synergies, to compensate for some possible flaws such as image quality impairment, while increasing the accuracy of the outcomes and bridging the gap towards generalization. Joint segmentation and registration models such as [16], [18] (joint phase field approximation and registration), [21] (model based on metric structure comparison), [15], [25] (level set formulation that merges the piecewise constant Mumford-Shah model with registration principles), [17] (grounded in the expectation maximization algorithm), [14] (based on a nonlocal characterization of weighted-total variation and nonlocal shape descriptors), or [1], [20], [23], [24], [27], fall within this framework.…”
“…So for almost every t ∈]0,T [, ψ k (t) strongly converges toψ(t) in W 1,4 (Ω, R 2 ) and det ∇ψ k (t) → k→+∞ det ∇ψ(t) in L 2 (Ω). From what was done in the stationary case [15], for almost every t ∈]0,T [,…”
Section: Theoretical Resultsmentioning
confidence: 99%
“…holds, these three elements combined allowing to handle the fidelity term. In order to deal with the nonlinearity in ∇ϕ, we propose introducing an auxiliary variable V i simulating the Jacobian deformation with a quadratic penalty method as in [15]. The decoupled problem becomes :…”
Section: Numerical Resolutionmentioning
confidence: 99%
“…We then use an alternating minimization scheme to solve the problem. We refer the reader to [15,Section 4.3. ] for an exhaustive description of the algorithm relying on the derivation of Euler-Lagrange equations solved by an L 2 gradient flow algorithm and an implicit/semi-implicit Euler time stepping scheme.…”
Section: Numerical Resolutionmentioning
confidence: 99%
“…Sharing representation between tasks and carefully intertwining them allows to reduce error propagation, to create synergies, to compensate for some possible flaws such as image quality impairment, while increasing the accuracy of the outcomes and bridging the gap towards generalization. Joint segmentation and registration models such as [16], [18] (joint phase field approximation and registration), [21] (model based on metric structure comparison), [15], [25] (level set formulation that merges the piecewise constant Mumford-Shah model with registration principles), [17] (grounded in the expectation maximization algorithm), [14] (based on a nonlocal characterization of weighted-total variation and nonlocal shape descriptors), or [1], [20], [23], [24], [27], fall within this framework.…”
“…In recent years, joint image processing models have experienced increasing attention, including combined segmentation/registration models [30,34] (joint phase field approximation and registration), [45] (model based on metric structure comparison), [26,61] (level set formulation that merges the piecewise constant Mumford-Shah model with registration principles), [33] (grounded in the expectation maximisation algorithm), [25] (based on a nonlocal characterisation of weighted-total variation and nonlocal shape descriptors), or [1,43,52,55,63,68]; joint image reconstruction and motion estimation [9,14,19,51,57,62,13,46,6]; joint reconstruction and registration for post-acquisition motion correction [22] with the goal to reconstruct a single motion-free corrected image and retrieve the physiological dynamics through the deformation maps, joint optical flow estimation with phase field segmentation of the flow field [12], or joint segmentation/optimal transport models [10] (to determine the velocity of blood flow in vascular structures). This can be attributed to several factors: (i) the will to limit error propagation.…”
In medical image analysis, constructing an atlas, i.e. a mean representative of an ensemble of images, is a critical task for practitioners to estimate variability of shapes inside a population, and to characterise and understand how structural shape changes have an impact on health. This involves identifying significant shape constituents of a set of images, a process called segmentation, and mapping this group of images to an unknown mean image, a task called registration, making a statistical analysis of the image population possible. To achieve this goal, we propose treating these operations jointly to leverage their positive mutual influence, in a hyperelasticity setting, by viewing the shapes to be matched as Ogden materials. The approach is complemented by novel hard constraints on the L ∞ norm of both the Jacobian and its inverse, ensuring that the deformation is a bi-Lipschitz homeomorphism. Segmentation is based on the Potts model, which allows for a partition into more than two regions, i.e. more than one shape. The connection to the registration problem is ensured by the dissimilarity measure that aims to align the segmented shapes. A representation of the deformation field in a linear space equipped with a scalar product is then computed in order to perform a geometry-driven Principal Component Analysis (PCA) and to extract the main modes of variations inside the image population. Theoretical results emphasizing the mathematical soundness of the model are provided, among which existence of minimisers, analysis of a numerical method of resolution, asymptotic results and a PCA analysis, as well as numerical simulations demonstrating the ability of the modeling to produce an atlas exhibiting sharp edges, high contrast and a consistent shape.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.