Vector Boosting for Rotation Invariant Multi-View Face Detection

In this paper, we propose a novel tree-structured multi-view face detector (MVFD), which adopts the coarse-to-fine strategy to divide the entire face space into smaller and smaller subspaces. For this purpose, a newly extended boosting algorithm named Vector Boosting is developed to train the predictors for the branching nodes of the tree that have multi-components outputs as vectors. Our MVFD covers a large range of the face space, say, +/-45? rotation in plane (RIP) and +/-90? rotation off plane (ROP), and achieves high accuracy and amazing speed (about 40 ms per frame on a 320?240 video sequence) compared with previous published works. As a result, by simply rotating the detector 90?, 180? and 270?, a rotation invariant (360? RIP) MVFD is implemented that achieves real time performance (11 fps on a 320? 240 video sequence) with high accuracy.

• Type: Conference
• Year: 2005
• Where: ICCV
• Authors: Chang Huang, Haizhou Ai, Yuan Li, Shihong Lao
• Keywords: Algorithm Named Vector, Computer Vision, Entire Face Space, ICCV 2005, Multi-view Face Detector, Smaller Subspaces, Video Sequence