Intelligent Technologies for Automated Electronic Systems

An Indagation of Biometric Recognition Through Modality Fusion

Author(s): P. Bhargavi Devi* and K. Sharmila

Pp: 125-133 (9)

DOI: 10.2174/9789815179514124010012

* (Excluding Mailing and Handling)

Abstract

One of the key predictions that had combined bio-sciences with innovation was bio-metrics, which represents a tool for security and criminology analysts to develop more accurate, robust, and certain frameworks. Biometrics, when combined with different combination techniques like feature-level, score-level, and choice-level combination procedures, remained one of the most researched technologies. Starting from uni-modular biometrics as unique marks, faces, and iris, they progress to multimodal bio-metrics. By presenting a similar investigation of frequently used and referred to uni- and multimodal biometrics, such as face, iris, finger vein, face and iris multimodal, face, unique mark, and finger vein multimodal, this paper will attempt to lay the groundwork for analysts interested in enhanced biometric frameworks. This comparative research includes the development of a comparison model based on DWT and IDWT. The method towards combining the modalities also entails applying a single-level, two-dimensional wavelet (DWT) that has been cemented using a Haar wavelet to accomplish the best pre-taking care of to eliminate disturbance. Each pixel in the picture is subjected to a different filtering operation in order to determine the peak signal to noise extent (PSNR). This PSNR analyses the mean square error (MSE) to quantify the disruption to hail before playing out the division of the largest dataset to the chosen MSE. In the most recent advancement, each pixel's concept is fixed up using the opposing two-dimensional Haar wavelet (IDWT), creating a longer image that is better able to recognise approbation, affirmation, and confirmation of parts. The MATLAB GUI is used to implement the diversions for this enhanced blend investigation, and the obtained outcomes are satisfactory.

© 2024 Bentham Science Publishers | Privacy Policy