Automated facial landmark detection, comparison and visualization

Abstract

Anthropometric facial landmarks and their detection has wide application in anthropology and forensic science. More specifically, these landmarks play an important role in facial comparison, in the analysis of morphological changes during human growth and in searching for the variability of human faces (e.g., sexual dimorphism). The automatic extraction of landmarks is a challenging area and specific visualization methods serving for the comparison are crucial for the correct interpretation of the results. In this paper we introduce a novel system which incorporates a newly proposed algorithm for automatic landmark detection. The algorithm is based on surface curvature and symmetric profile extraction. The landmarks are then used in the Generalized Procrustes analysis for facial comparison. Our system also introduces appropriate visualization methods supporting the presentation and understandability of the results.