Nose detection and face extraction from 3D raw facial surface based on mesh quality assessment

We propose a novel approach for nose tip detection and extraction of frontal face area from raw 3D triangular mesh facial surfaces. Our approach is inspired from the observation that the region around some facial landmarks are characterized by low mesh quality. To measure and assess the quality of the mesh surface, we present an original framework using which we extract a group of candidate triangular facets. In the second stage of our algorithm, we find the single facet that corresponds to the nose tip from the group of candidate triangle facets using a series of filtering steps. In addition to exploring the performance of our model, we also show how the mesh quality assessment framework can be generalized to extract the frontal face area from any raw face scan data. Tests conducted on a group of 240 raw facial scans prove the robustness of our method, and do not show cases of failure.