Efficient Searching Technique for Biometric Data Based on Inverted Files

Abstract

     The increase in the amount of biometric data in databases has caused problems for efficient data retrieval and management. Biometric databases face challenges primarily due to the high dimensional nature of biometric features, such as fingerprints, facial recognition data, and iris scans, leading to time consumption in retrieval and matching as well as computationally intensive, and inefficiencies that can hinder real-time applications, especially when using multiple biometric traits. To address these issues, this paper presents a mechanism for searching three-dimensional biometric databases based on inverted files specifically tailored to handle biometric data. As well as designing a 3D database to store data. This method was applied to the proposed multi-biometric recognition system using iris and fingerprint features by storing the biometric data in the database based on identifiers, that are retrieved from the inverted files to locate the vectors of the candidates and retrieve them to perform the matching, thus improve the matching speed without the need to search and match the entire feature vectors in the database. The results showed that less time is consumed in retrieving and matching when applying this technique compared to the traditional method.