A framework for generating anatomically detailed subject-specific human facial models for biomechanical simulations

Realistic biomechanical simulations of the human face rely on detailed and accurate anatomical models. A recent study involving ultrasound imaging revealed that tissue structures in the human face can be separated into two major strata that move independent of each other. Based on this observation, anatomically accurate finite element models representing soft tissues in both layers of the human face were developed using 3D segmented data derived from the high-resolution US Visible Human cryosection images. The three-dimensional geometry of these tissue structures was described using Cubic Hermite finite elements. The use of Hermite family elements ensures the continuity of displacement gradient across element boundaries and hence maintains the moment balance throughout the computational domain in mechanical simulations. This in turn leads to more accurate predictions of soft tissue deformations. Creating subject-specific detailed model of the face suitable for biomechanical analysis is, however, a time-consuming task. This paper proposes a fast semi-automated framework for generating detailed subject-specific facial models including internal muscles using techniques that involve landmark-based affine transformation, iterative surface-fitting and free-form deformation. Generated models for three individuals are presented to demonstrate the efficacy of the proposed methodology.