Title :
Using Bilevel Feature Extractors to Reduce Dimensionality in Images
Author :
Joliveau, Marc ; Gendreau, Michel
Author_Institution :
Univ. de Montreal, Montréal, QC, Canada
Abstract :
A bilevel procedure for dimensionality reduction makes it possible to discover the underlying global geometry of a complex natural observations dataset-such as human handwriting or faces under different viewing positions-with higher precision than existing methods.
Keywords :
feature extraction; geometry; bilevel feature extractor; complex natural observation dataset; dimensionality reduction; human handwriting; Database systems; Feature extraction; Fingerprint recognition; Laplace equations; Lighting; Principal component analysis; Visual perception; dimensionality reduction; feature extractor; image processing; pattern recognition; scientific computing;
Journal_Title :
Computing in Science & Engineering
DOI :
10.1109/MCSE.2011.55
