Iterative unsupervised object detection system
S. Onis, Henri Sanson, Christophe García · 2008
Current object detection systems provide good results, at the expense of requiring a large training database. This paper presents an unsupervised iterative object detection system using a selection of previously detected objects in order to perform new object detection. Our experiments show that this method enables face detection with a greatly reduced set of examples and outperforms the detection rates of our non iterative detection system based on normalized cross-correlation and affine deformation compensation.