Department of Image Processing

While the problem of facial landmark detection has been recently getting a big attention in the computer vision community, most of the methods deal only with near-frontal views and there are only a few really multi-view detectors available. In this work, we tackle the problem of designing a multi-view facial landmark detector which is robust and works in real-time on low-end hardware. Our landmark detector is an instance of the structured output classifiers describing the face by a mixture of tree-based Deformable Part Models (DPM). We propose to learn parameters of the detector by the Structured Output Support Vector Machine algorithm which, in contrast to existing methods, directly optimizes a loss function closely related to the standard evaluation metrics used in landmark detection.