Detection of Grey Human Face Based on Digital Elevation Models

Based on the shape of landscape, a practical algorithm for detecting human face is put forward. This method takes a human face as a special kind of landscape, so digital elevation models (DEM) could be employed to detect candidate human faces. We first extract features of human faces by means of three kind of curvatures of faces (general Gaussian curvature, section curvatures and tangent curvatures) based on DEM, and find all possible areas of human faces by feature matching, then determine correct human faces by further using the so called "lakes" and "mountains". Experiments show this algorithm is effective and could be applied in practice.