An Approach of Combining Iris and Fingerprint Biometric At Image Level in Multimodal Biometrics System
S. M. Rajbhoj1, P. B. Mane2
1S. M. Rajbhoj, Ph.D. Research Scholar, BVUCOE, Bharati Vidyapeeth University, Pune, India.
2Dr. P. B. Mane, Principal, AISSMS, Institute of Information Technology Pune, India.
Manuscript received on February 20, 2015. | Revised Manuscript received on February 28, 2015. | Manuscript published on March 05, 2015. | PP: 102-106 | Volume-5 Issue-1, March 2015. | Retrieval Number: A2545035115/2015©BEIESP
Open Access | Ethics and Policies | Cite
©The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/
Abstract: Biometric systems depending on single source of information has many limitations. These are noisy input data, inability to enroll, unacceptable error rates, universality of traits and spoofing. Multimodal biometric system overcomes these limitations by combining information from multiple sensors. In Image fusion usually images are extracted from single trait using different sensors. This type of fusion is generally used when feature set are homogenous. In this paper a multibiometric system using image level fusion of two most used biometric traits, fingerprint and iris is proposed. The feature set obtained from iris and fingerprint images are incompatible, non-homogenous and relationship between them is not known. Here the pixel information is fused at image or feature level. A unique feature vector is constructed from the textural information of fused image of fingerprint and iris. Feature vector is stored as template and used for matching. Matching is carried using Hamming distance. The proposed framework is evaluated using standard database and database created by us. The system overcomes limitation of unimodal biometric system and equal error rate of 0.4573 has been achieved.
Keywords: biometric, fingerprint, iris; wavelet transform, texture, feature level, fusion, hamming distance.