Qifa Ke, Gang Xu, and Song De Ma January 1998
In Proceedings of the Sixth International Conference on Computer Vision (ICCV)