Enhanced Pictorial Structures for Precise Eye Localization under Uncontrolled Conditions

In this paper, we present an enhanced Pictorial Struc- ture (PS) model for precise eye localization, a fundamen- tal problem involved in many face processing tasks. PS is a computationally efficient framework for part-based ob- ject modelling. For face images taken under uncontrolled conditions, however, the traditional PS model is not flexi- ble enough for handling the complicated appearance and structural variations. To extend PS, we 1) propose a dis- criminative PS model for a more accurate part localization when appearance changes seriously, 2) introduce a series of global constraints to improve the robustness against scale, rotation and translation, and 3) adopt a heuristic predic- tion method to address the difficulty of eye localization with partial occlusion. Experimental results on the challenging LFW (Labeled Face in the Wild) database show that our model can locate eyes accurately and efficiently under a broad range of uncontrolled variations involving poses,...

• Type: Conference
• Year: 2009
• Where: CVPR
• Authors: Xiaoyang Tan (Nanjing University of Aeronautics and Astronautics), Fengyi Song (Nanjing University of Aeronautics and Astronautics), Zhi-Hua Zhou (Nanjing University), Songcan Chen (Nanjing University of Aeronautics and Astronautics)
• Keywords: Computer Vision, CVPR 2009, Eye Localization, Pictorial Structures