Linear and Nonlinear Dimensionality Reduction Techniques

Ioan Buciu

Abstract

This overview aims at describing the most representative existing standard as well as new linear and nonlinear data dimensionality reduction methods. Finding new dimensionality reduction algorithms for preserving as much as possible the underlying structure of data are still challenging topics from the computer vision field. The paper presents the most classical and advanced techniques used for data dimensionality reduction.

Keywords

dimensionality reduction, image processing, computer vision, pattern recognition, artificial intelligence.

View full article