A robust facial landmark detection method in multi-views

Xinran Liu, Fei Su

Abstract: Facial landmark detection is a vital step in face recognition. Current facial landmark detection methods have not obtained the satisfied results to handle multi-view face images. In this paper, we propose a Multi-ESR method based on Explicit Shape Regression method, in which different ESR models are trained in different views. Furthermore, a simple and effective method to estimate the head pose is also described, the estimation result is used in Multi-ESR to choose the corresponding model. Experimental results show the proposed method yielding better performance comparing to the state-of-art methods handling multi-view face images.