Plotting and reading accurate facial maps
The 'Spectral mesh processing for craniofacial dysmorphology' (SP-MORPH)(opens in new window) project has developed techniques for facial analysis based on spectral decomposition methods. Spectral mesh processing (SMP) has enabled analysis of the whole facial surface in contrast to traditional methods that relied on a reduced set of landmark points. Data pre-processing involved verification of the composition and quality of the available datasets. The researchers successfully established a robust automated landmarking algorithm. Furthermore, they introduced additional processing that eliminated some of the undesirable artefacts such as holes and disconnected parts. They developed a framework for quantitative evaluation of hole-filling algorithms and generated a realistic dataset of synthetic patches that served as benchmark material. Normalisation involved application of 'Least Squares Conformal Maps'. Appropriate re-sampling and inverse mappings from 2D to 3D enabled a new representation of the input surfaces in which the whole set is in correspondence.
