An automatic approach for 3-D facial shape change analysis by combination of ASM and morphometric tools

Abstract

This paper presents a new approach for automatic facial shape change analysis by combining a stereo-assisted active shape model (ASM) with morphometric tools. We firstly describe how to extend the 2-D ASM to 3-D space by stereo correspondence search and disparity map interpolation for automatic facial landmark extraction. Morphometric techniques such as generalized Procrustes analysis, principal component analysis and thin plate spline decomposition are then reviewed and applied to the quantitative analysis of shape changes in 2-D facial midline profiles and 3-D facial landmarks. The proposed method is validated both statistically and visually by characterizing bite-block induced shape changes in a heterogeneous sample of young orthodontic patients.