A Guided Tour Of Statistical Shape Models

A statistical shape model is a concise depiction of a cluster of shapes. This representation is based on Principal Component Analysis (PCA) performed on this cluster. Principal Component Analysis (PCA) aims to reduce the size of the model by selectively retaining eigenvectors that capture most of the original variability. The statistical model generates shapes that are geometrically consistent with the shapes assimilated by the model, achieved by varying a single parameter influencing the expression of the eigenvalues encoding the variability.