Independent comparative study of PCA, ICA, and LDA on the FERET data set

Abstract

Face recognition is one of the most successful applications of image analysis and understanding and has gained much attention in recent years. Various algorithms were proposed and research groups across the world reported different and often contradictory results when comparing them. The aim of this paper is to present an independent, comparative study of three most popular appearance-based face recognition projection methods (PCA, ICA, and LDA) in completely equal working conditions regarding preprocessing and algorithm implementation. We are motivated by the lack of direct and detailed independent comparisons of all possible algorithm implementations (e.g., all projection-metric combinations) in available literature. For consistency with other studies, FERET data set is used with its standard tests (gallery and probe sets). Our results show that no particular projection-metric combination is the best across all standard FERET tests and the choice of appropriate projection-metric combination can only be made for a specific task. Our results are compared to other available studies and some discrepancies are pointed out. As an additional contribution, we also introduce our new idea of hypothesis testing across all ranks when comparing performance results.

What's hot and what's not: Windowed developer topic analysis

“…Grant et al [5] and have used an alternative technique, called Independent Component Analysis [3] to separate topic signals from software code.…”

Personal identification using Eigenfeet, Ballprint and Foot geometry biometrics

“…Similar to Eigenfaces suggested by Turk and Pentland [7] for face recognition this Eigenfeet matcher is sensitive to both geometrical and textural properties. For a probe set of 1195 facial images of subjects taken at the same time Delac et al [15] could verify PCAbased recognition rates of 82.26% at rank one, i.e. within the top one match.…”

Towards the Low False Alarms and High Detection Rate in Intrusions Detection System

“…They used PCA (Principal Component Analysis) and GA for preprocessing and MLP for features classification using KDD-cup dataset. LDA outperforms PCA and PCA is not suitable for large dataset [4], hence their work is limited for small size datasets and results are not more realistic to actual network traffic as there are approved deficiencies in KDD-Cup dataset.…”

“…There has been a tendency to use PCA approach for features subset selection or reduction in many different domains like face recognition, image compression as well as intrusion detection [10] but LDA has more benefits over PCA and is preferred over PCA due to the following reasons. 1) LDA outperforms PCA in case of large dataset [4]. 2) LDA directly deals with both discrimination withinclasses as well as between-classes while PCA does not have any concept of the between-classes structure [1].…”

“…which analogous to the extraordinary ability of the human mind to reason and learn in an environment of uncertainty and imprecision [4]. Selection of suitable dataset is a backbone of efficient intrusion detection approaches.…”