Fast search methods for biometric signal processing

The aim of this paper is to investigate some current biometric algorithms - which are basically statistical pattern matching and pattern database searching- and ways of speeding them up. This will encompass both optimal (where the match found is identical to an exhaustive search) and suboptimal techniques. Implementation of the fast-search algorithms investigated to determine both the computational complexity and optimality of the search performance. The fast evolution and increasing complexity of computing platforms poses a major challenge for developers of fast algorithms for signal processing. Even though processor speeds have increased dramatically, the everincreasing complexity of algorithms being investigated for signal processing applications still places constraints on processing times. This is particularly evident in the area of biometric signal processing. Airport security is a major issue after the event of 9/11. Now countries have started issuing biometric passports or dasiasmartpsila passports, which are basically an embedded chip, containing information of the identity of the person. There are many biometric identification techniques currently under investigation, including voice, face, and iris, fingerprint, and signature identification. Each technique has its strength and some limitations. One problem with countries issuing biometric passports or dasiasmartpsila passports is that even if the algorithms work well (which they donpsilat always), the speed of processing is a problem in many application domains (for example, Airports where large numbers of individuals need to be screened quickly but with a high degree of accuracy). Code for vector quantization and fast vector quantization search methods was developed using MATLAB.