2001 Volume 21 Issue 1Supplement Pages 187-190
In this paper we will describe some of the research done in using wavelet transforms for pattern recognition at the Centre for Research in Pattern Recognition and Machine Intelligence at Concordia University. This includes road signs recognition, face detection and recognition, oriental characters recognition, hand-written numerals and characters, image compression, and image de-noising, etc. We have developed a new metric called the Hellinger-Kakutani metric for measuring the distance between two descriptors for classification. This metric is filter-invariant, hence can be used on noisy images.