Improving Biometric Authentication Performance From the User Quality

The effectiveness of a biometric measurement and sensing system is directly related to the performance generated from sensed data. This paper investigates a new approach to quantify the quality of sensed data from the user templates. The objective is to incorporate the quality of sensed data to generate a reliable estimate on the matching scores. The proposed method of extracting user quality is based on the confidence of generating reliable matching scores from the user templates. We simultaneously extract the palmprint and hand-shape images from the single hand image and ascertain the performance improvement for the individual trait. The experimental results from the proposed approach are also presented when the biometric measurements from the finger knuckles are employed. The experimental results presented in this paper show significant improvement in performance while incorporating the proposed method of user quality in the matching stages. The proposed user-quality-based fusion of the two biometric modalities also achieves promising improvement in performance.