Biometric Recognition Systems - Then and Now

Biometrics have been in existence from very early times. Fingerprints were used as a non-counterfeitable mark as far back as 500 BC.

Babylonian merchants used fingerprints to secure business transactions. The undertakings were recorded in clay tablets with added fingerprints. Footprints and Fingerprints have been historically employed to distinguish children.  Inked fingerprints of children were also taken for identification purpose by Chinese merchants during the 14th century and early Egyptians differentiated between traders by various biometric-type attributes.

Bertillon has been credited with the systematic study of the measurement of human beings. The system developed by him (anthropometry) was used in fighting crime. Francis Galton developed a classification system for fingerprints. By 1936, the concept of using iris pattern for identification was proposed.

Later, the precursors of modern voice recognition systems were developed. Similarly, iris recognition, signature recognition and hand geometry biometric devices were developed. However, fingerprint recognition ruled the biometric market and would have continued to do so, had not semi-automated face recognition system made its appearance in 1960s.

In a (1970s) attempt to automate a recognition system, Goldstein, Harmon and Lesk used 21 specific markers on the face. However, the measurements and locations on the face were manually computed. In 1988, Kirby and Sirovich applied algebra techniques to it for more accurate results. This was a landmark achievement in Biometric face recognition system.

Modern automated facial detection applications began in 1991. This technology was further developed for security purposes.

There are two mainstream approaches - Geometric (feature based) and Photometric (view based). Many algorithms were developed in this technology.

Three main ones among them are:

* PCA: Principle Components Analysis (PCA) is an approach, where the probe and gallery images must be normalized to line up the eyes and mouth of the subjects within the images.

* LDA: Linear Discriminate Analysis (LDA) is a statistical approach for classifying samples of unknown classes based on training samples with known classes.

* EBGM: Elastic Bunch Graph Matching (EGBM) is an approach, where the non-linear characteristics that are not addressed by the linear analysis methods, are measured for recognition of a face.

Modern biometric face recognition systems rely on these algorithms for identification.

The computerized technology has made considerable progress in recent years. It is now widely used in surveillance for security purposes. Wanted criminals, suspected terrorists and missing people can be detected with the help of this science. Face recognition is used for public scrutiny in airports, hospitals, schools and other places where crowds gather.

Biometric face recognition systems are a booming technology, with utilization rapidly increasing.

However, Civil rights activists are generally opposed to its widespread usage for reasons related to loss of privacy. It is strongly asserted that behavior of people may be unknowingly recorded and abused later. However, its promising role in spotting illegitimate behavior cannot be ignored. The software can be used in combination with Closed-circuit Surveillance Cameras (CCTVs) for smarter surveillance. Solution providers for this system have seen an upsurge.