Lightweight Self-Driven Deformable Organ Animations

TL;DR

This work tackles the challenge of producing interactive, real-time visualizations of large deformable internal organs for medical training. It introduces a lightweight, self-driven deformable model that uses region-based partitioning and stochastic tuning to generate coordinated oscillatory motions, driven by internal constraints rather than external inputs. The key contributions include scalable mesh partitioning for performance/accuracy tradeoffs and a workflow that seeds physics-based animation with pre-recorded motion data for rapid prototyping. The approach offers practical benefits for medical education and visualization, with potential extensions to inter-organ coupling and fluid dynamics for more realistic simulations.

Abstract

The subject of simulating internal organs is a valuable and important topic of research to multiple fields from medical analysis to education and training. This paper presents a solution that utilizes a graphical technique in combination with a Stochastic method for tuning an active physics-based model. We generate responsive interactive organ animations with regional properties (i.e., areas of the model oscillating with different harmonic frequencies) to reproduce and capture real-world characteristics. Our method builds upon biological and physical discoveries to procedurally generate internally controlled rhythmic motions but also enable the solution to be interactive and adaptive. We briefly review deformation models for medical simulations and investigate the impediments to combining 'computergraphics' representations with biomechanical models. Finally, we present a lightweight solution that is scalable and able to procedurally generate large organ animations. In particular, simplified geometric representations of deformable structures that use periodic coupled forces to drive themselves.

Figures (7)

• Figure 1: Timeline - Visual illustration of graphical and interactive related publications over the past few years that have contributed towards more active and engaging soft-body systems. [A] kim2012physics, [B] georgii2005interactive, [C] cheney2013unshackling, [D] kenwright2014planar, [E] tan2012soft, [F] tan2011articulated, [G] rieffel2014growing, [H] lehman2011evolving, [I] shim2003generating, [J] stavness2014muscle, [K] BKKS.
• Figure 2: Categorising - Typical areas of investigation for achieving realistic soft-tissue models (static and dynamic).
• Figure 3: Overview - Interconnected elements to construct and tune the soft-body's self-driven motions.
• Figure 4: Lungs - Three-dimensional mesh model of the organ (e.g., lung), classify different regions of the mesh (small sub-areas), oscillating motion based on the contraction/expansion of tuned constraints.
• Figure 5: Heart - Three-dimensional mesh model of the organ (e.g., heart), classify different regions of the mesh (small sub-areas), oscillating motions based on the contraction/expansion of tuned constraints.