Image sensor fusion for multimodal biometric recognition in mobile devices

Abstract

Aim

Multimodal biometric recognition authenticates or identifies someone using various biometric features or modalities. Multimodal biometric systems use different biometric qualities to improve the recognition process's accuracy, security, and reliability rather than depending solely on one biometric property, such as fingerprint or facial recognition. Sensor fusion techniques can merge data from several biometric sensors, including fingerprint scanners and facial recognition cameras, to boost the accuracy and dependability of biometric authentication in the context of multimodal biometric recognition in mobile devices. The process of combining data from numerous image sensors to enhance a system's overall performance is known as image sensor fusion.

Challenges

Challenges includeintegration difficulties in achieving accurate and secure multimodal biometric recognition on mobile devices. The study goal was Image sensor fusion for multimodal biometric recognition in mobile devices.

Methodology

The study uses a specially made database of 450 face photos roughly 450 fingerprints.Weiner filter (WF) is used for preparing data. Battle Royale optimized with deep convolutional neural network (BRO-DCNN) is proposed to improve the overall performance of fingerprint and facial information.

Findings

Accuracy, sensitivity, specificity, and precision are used to measure the effectiveness of the proposed BRO-DCNN system. Study results show that BRO-MLANN solves the many characteristics of smart cities and is an efficient approach compared to other current methods in recognition rate, 98 % accuracy, 99 % precision, 96 % specificity, and 94 % sensitivity.