|International Journal of Applied Information Systems|
|Foundation of Computer Science (FCS), NY, USA|
|Volume 10 - Number 10|
|Year of Publication: 2016|
|Authors: Oriola Oluwafemi, Orimoloye Segun Michael|
Oriola Oluwafemi, Orimoloye Segun Michael . Quad-search: Modified Binary Search Algorithm for Efficient Fingerprint Biometrics Identification in Multicore Architecture. International Journal of Applied Information Systems. 10, 10 ( May 2016), 28-32. DOI=10.5120/ijais2016451559
Linear Search Algorithm is often applied to Fingerprint Biometrics Identification in traditional computing. With the advent of multicore systems, more attention must be placed on Binary Search Algorithms for efficient Fingerprint Biometric Identification in large databases. However, Binary Search Algorithms can conveniently be parallelized for two cores. Therefore, this study assumes that a variant of Binary Search Algorithm, which supports multi-parallelization, will take better advantage of advances in multicore systems and improve efficiency of Biometric Identification. This paper hence presents a Modified Binary Search algorithm known as Quad-search for Biometric Identification. The Fingerprint Identification Algorithm is developed by splitting the fingerprint array at the midpoint and parallelizing the search in each split to form four sub-arrays. To evaluate the proposed algorithm, a real-life Fingerprint Biometric Identification System of tested fingerprint images is used. The execution time of the Modified Binary Search Algorithm is compared with both Binary Search and Linear Search Algorithms in an Intel Quadcore Architecture. The results show that at near, middle, and far fingerprint positions in the database, the Modified Search Algorithm is executed at lower time compared to both the Parallelized Binary Search Algorithm and Linear Search Algorithm, but the parallelized Binary Search Algorithm executes at lower time at mid fingerprints locations compared to Linear Search Algorithm. The running time is estimated as O(Log(logN)) The Modified Binary Search Algorithm resulted in efficient Fingerprint Biometrics Identification in multicore architecture.