Efficient fingerprint search based on database clustering

  • Authors:

  • Manhua Liu;Xudong Jiang;Alex Chichung Kot

  • Affiliations:

  • EEE, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore;EEE, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore;EEE, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

  • Venue:
  • Pattern Recognition
  • Year:
  • 2007

Quantified Score

Hi-index 0.01

Visualization

Abstract

Fingerprint identification has been a great challenge due to its complex search of database. This paper proposes an efficient fingerprint search algorithm based on database clustering, which narrows down the search space of fine matching. Fingerprint is non-uniformly partitioned by a circular tessellation to compute a multi-scale orientation field as the main search feature. The average ridge distance is employed as an auxiliary feature. A modified K-means clustering technique is proposed to partition the orientation feature space into clusters. Based on the database clustering, a hierarchical query processing is proposed to facilitate an efficient fingerprint search, which not only greatly speeds up the search process but also improves the retrieval accuracy. The experimental results show the effectiveness and superiority of the proposed fingerprint search algorithm.