Fast Musculoskeletal Registration Based on Shape Matching

Abstract

Abstract. This paper presents a new method for computing elastic and plastic deformations in the context of discrete deformable model-based registration. Internal forces are estimated by averaging local transforms between reference and current particle positions. Our technique can accommodate large non-linear deformations, and is unconditionally stable. Moreover, it is simple to implement and versatile. We show how to tune model stiffness and computational cost, which is important for efficient registration, and demonstrate our technique in the complex problem of inter-patient musculoskeletal registration.

Computer-assisted analysis of human upper arm flexion by 4D-visualization based on MRI

“…Finally, polygonal models of the muscles may serve as input for various kind of musculoskeletal research from biomedical engineering [6]. Sturmat et al used these models for Finite Element analysis of the contraction of the biceps brachii muscle [15].…”

“…Notably, the results were surface models representing real time contour deformations. In [6], a general framework aimed at biomedical engineering for macroscopic 4D-modeling of skeletal muscles is given, especially for the human gastrocnemius and the biceps brachii muscle.…”

Minimally Interactive MRI Segmentation for Subject-Specific Modelling of the Tongue

“…Somphone et al transformed their binary template, subject to conformity constraints between local patches [25]. In a different approach, Gilles et al used explicit shape representation of the template to segment the musculoskeletal structures out of MR images [26]. The mesh deformation was regularized based on an expanded version of a computer animation technique called shape matching [27].…”

Manifold-enhanced Segmentation through Random Walks on Linear Subspace Priors

“…In [6,7], segmentation is achieved with multi-object deformable models. Deformable models ( [11,12]) are surface models superimposed to the c 2011.…”

“…target image in an initial state and deformed by minimizing an objective function composed of two terms: a data-term -attracting the surface to detected image contours -and a regularization term, ensuring the surface remains realistically smooth or similar enough to a reference surface (as in [7]). In [4], prior knowledge on the shape of one muscle is further enforced by imposing the model to reside in a hierarchical shape space built via diffusion wavelets decomposition on training examples.…”