Correlation and Independent Component Analysis Based Approaches for Biometric Recognition
Résumé
Independent component analysis (ICA) models, describing a given signal as a linear
combination of various independent sources, have proven to be a fruitful endeavor. One
prominent example deals with audio applications in order to separate the speaker’s voice
from environmental noises disturbing it. However, very few ICA based systems are
available for biometric encryption applications. For that specific purpose, the ICA
method can be easily adapted to add noise to a target image in order to encrypt it. In this
chapter, at first, we discuss biometric recognition systems based on the ICA and
correlation approaches. Next, we explore an ICA-based algorithm for face recognition.
Basically, it consists of building a base of independent components using a learning
database that contains several chosen reference images. Then, the target image (image to
be recognized) is projected on the independent component base, and the similarity
between the target image and each of the reference images is studied. Discrimination tests
between the proposed technique and alternate methods are conducted by using the
Pointing Head Pose Image Database (PHPID). In this chapter we report some of the
recent developments dealing with the ICA method for face recognition applications. As
part of our analysis, we precisely determine a set of metrics aimed at better understanding
the role of the number and choice of the reference images on the performance of the
proposed technique.