Fast, Illumination Insensitive Face Detection Based on Multilinear Techniques and Curvature Features

This paper brings together two recent developments in image analysis. We consider a new mathematical framework that provides illumination invariant descriptors for face detection. Towards fast learning and processing, we understand images and the corresponding feature maps as multilinear entities and apply higher order classifiers for image analysis and object detection. Experimental results underline that this approach indeed provides quick training, fast runtime and robust performance across a variety of illumination conditions