Review of Dimensionality Reduction Techniques for Transformational Image Signals

Abstract

In recent years, graphs have become a primary method of representation of large data sets generated through ubiquitous sensing devices and there is a need for dimensionality reduction techniques to perform processing, analysis and storing of these large data sets in the graph setting effectively. However finding suitable dimensionality algorithms, for a particular application has become cumbersome and even more difficult for real-time applications concerned. So in this paper, we have attempted to review dimensionality techniques proposed for processing, analyzing the large data sets in the graph setting. We have kept our review based on the computational complexity of dimensionality reduction techniques.