Multimodal Biometrics Research Articles

Versatile Multimodal Biometrics to Prevent ATM Identity Frauds

—Automated Teller Machine (ATM) frauds have increased over the years resulting in the loss of customers and trust in current ATM security measures. In order to fight identity fraud and regain trust and reliability, biometric recognition technologies have become a necessity for ATMs. Using a card and pin is no longer reliable due to the rise of ATM identity fraud over the years. Studies have been carried out to solve this uprising issue, such as replacing the memorized pin with a one-time pin (OTP) and others using biometrics. There are still many gaps and challenges in the systems, with a common challenge being their accuracy rate, dependencies, and reliability. This study is conducted to alleviate the problem by introducing a versatile multimodal biometrics system. After a careful review of different biometric techniques and their differences in human identification inboth accuracy and reliability, physical biometrics isadopted in this study. Unlike behaviouralbiometrics, physical biometric features cannot easily be copied, replicated, or mimicked, and they are unique per individual. The three body features chosen for this study are the face, fingerprint, and iris. A real-time feature quality evaluation method is employed to assess the reliability of the biometric recognition results. The system provides independent anytime ATM access and prevents card-based theft and fraud, it can be relied on to always deliver uninterrupted ATM services to customers. Customers register their fingerprint, face, and iris to be able to use the biometric ATMs. A combination of any two of the registered biometric features can be used to authenticate users with high accuracy and reliability on ATMs with a 0% false acceptance rate. It has been found that allowing multiple options for users reduces false rejections and provides a 100% ATM access guarantee.

APPLICATIONS OF MULTIMODAL BIOMETRICS AUTHENTICATION FOR ENHANCING THE IOT SECURITY USING DEEP LEARNING

The Internet of Things (IoT) integrates billions of electronic devices into computer networks to provide advanced and intelligent services that enable devices to communicate with each other by exchanging information with minimal human interaction. The security issue is at higher risk in IoT systems than in other computing systems. Maintaining the security requirement when attacking the physical surface of the IoT system device is a challenging task. Implementing security mechanisms like authentication and access control for the IoT ecosystem is necessarily needed to ensure the security of IoT devices. The key used for security may be stolen, forgotten, or forged. Also, the key may be generated by intruders or men in the middle of traditional security mechanisms. Biometric security is becoming more advanced and sophisticated with technological advancements and is mostly used in authentication systems. In unimodal biometrics, only one biometrics character can be applied which does not apply to ensure the security of IoT systems. In this paper, Multimodal biometrics authentication was used for securing edge devices in the IoT ecosystems. Face image and fingerprint image were used as multimodal biometrics systems for authenticating users to secured IoT devices. A pi-camera module and fingerprint sensor were used to capture biometric data. Image processing techniques were then applied to the images. Then CNN algorithms were used for feature extraction and model creation. During model creation, the RELU function was used as an activation function, soft-max for image classification, and Max-pooling for image dimensional reduction which helped the model speed up the training process. Experimental results show that the accuracy of the face image and fingerprint image is 92% and 89% respectively, which is a promising result that achieves the objective of the study.
 Keywords: Internet of Things, Multimodal Biometrics, Authentication, CNN, Deep Learning

Multimodal biometrics authentication using extreme learning machine with feature reduction by adaptive particle swarm optimization

Multimodal biometrics authentication using extreme learning machine with feature reduction by adaptive particle swarm optimization

Multimodal Biometrics Authentication in Healthcare Using Improved Convolution Deep Learning Model

In healthcare applications, biometric authentication is crucial in managing patient credential details. The limited usage of biometric traits causes personal details theft, treatment hacking, and payment hijacking. Multimodal biometrics should incorporate into the healthcare system to maintain security and privacy in healthcare applications. Previous methods ensure authentication and security consume high computation complexity and fail to maintain system reliability and scalability. Therefore, in this work, an improved convolution deep learning model (ICDLM) is applied to manage the patient traits in the healthcare center. Initially, Experimentation is conducted with SWAN DATASET, which consists of face, voice, and periocular biometric features. The biometric characteristics are investigated according to the super point convolution neural model. The model analyzes the trait descriptors and interest points to identify the patterns. The extracted patterns are utilized to match the patient while accessing the information in healthcare applications. This process uses template pattern matching (TPM), which minimizes the absolute difference while authenticating the users. The system’s efficiency implemented using Python and excellency is compared with traditional approaches. The system attains 98.31% and 98.67% accuracy on the number of users and attempts with a minimum false identification rate of 0.0125 and 0.0128 for several users and attempts.

Efficient Multimodal Biometric Recognition for Secure Authentication Based on Deep Learning Approach

Biometric identification technology has become increasingly common in our daily lives as the requirement for information protection and control legislation has grown around the world. The unimodal biometric systems use only biometric traits to authenticate the user which is trustworthy but it possesses various limitations such as susceptibility to attacks, noise occurring in a dataset, non-universality challenges, etc. Multimodal biometrics technology has the potential to avoid the various fundamental constraints of unimodal biometric systems and also it has garnered interest and popularity in this respect. In this research, an efficient multimodal biometric recognition system based on a deep learning approach is proposed. The structure is implemented around convolutional neural networks (CNN) in which feature extraction and Softmax classifier are used to identify images. This method employs three CNN models for iris, face, and fingerprint were integrated to create the system. The two levels of fusion strategy such as feature level fusion and score level fusion were employed. The efficiency of the proposed model is evaluated based on the two most popular multimodal datasets as SDUMLA-HMT and BiosecureID biometric dataset. The result analysis demonstrates that the proposed multimodal biometric recognition provides the enhanced result with higher accuracy of 99.92%, a lower equal error rate of 0.10% on feature level, and 0.08% on score level fusion. Similarly, the average FAR is 0.09% and the average FRR is 0.06%. Because of this enhanced result, the proposed approach is computationally efficient.

Optimal Score Level Fusion for Multi-Modal Biometric System with Optimised Deep Ensemble Technique

ABSTRACT Multi-modal biometric refers to the use of various biometric indicators for individual identification by personal recognition systems. When compared to unimodal biometrics, which uses only one biometric data, such as a fingerprint, face, palm print or iris, multi-modal authentication offers a higher degree of authentication. A new optimal score value that fuses deep learning and multi-modal biometrics would be produced in the project study. A proposed approach was split into three main groups: feature extraction, pre-processing and ensemble recognition. First, median filtering and ROI procedures were utilised for pre-processing-captured original biometric information for the wrist, dorsal, palm vein and palm print. Pertinent features are then retrieved from the corresponding preprocessed images for every modality. These extracted features are subjected to an imposter or genuine determination. Neural Network 1, Neural Network 2 and Deep Convolution Neural Network create a new deep ensemble model in the event of a forgery or accurate estimation (DCNN). The outputs of NN1 and NN2 are the inputs to DCNN, which provides information on whether the biometric data are authentic or not. Finally, the results are fine-tuned by the weight of DCNN utilising new hybrid optimisation scheme referred as Butterfly combined Tunicate Swarm Optimisation (BTSA).

Multi-modal Biometrics Based Implicit Driver Identification System Using Multi-TF Images of ECG and EMG

As security is emphasized inside and outside the vehicle, research on driver identification technology using bio-signals is being actively studied. The bio-signals acquired by the behavioral characteristics of the driver include artifacts generated according to the driving environment, which could potentially degrade the accuracy of the identification system. Existing driver identification systems either remove the normalization process of bio-signals in the preprocessing stage or use artifacts included in a single bio-signals, resulting in low identification accuracy. To solve these problems in a real situation, we propose a driver identification system that converts ECG and EMG signals obtained from different driving conditions into 2D spectrograms through multi-TF image and uses multi-stream CNN. The proposed system consists of a preprocessing phase of ECG and EMG signals, a multi-TF image conversion process, and a driver identification stage using a multi-stream-based CNN. Under all driving conditions, the driver identification system reached an average accuracy of 96.8% and an F1 score of 0.973, which overperformed the existing driver identification systems by more than 1%.

FIMBISAE: A Multimodal Biometric Secured Data Access Framework for Internet of Medical Things Ecosystem

Information from the Internet of Medical Things (IoMT) domain demands building safeguards against illegitimate access and identification. Existing user identification schemes suffer from challenges in detecting impersonation attacks which leave systems vulnerable and susceptible to misuse. Significant advancement has been achieved in the domain of biometrics and health informatics. This can take a step ahead with the usage of multimodal biometrics for the identification of healthcare system users. With this aim, the proposed work explores the fingerprint and iris modality to develop a multimodal biometric data identification and access control system for the healthcare ecosystem. In the proposed approach, minutiae-based fingerprint features and a combination of local and global iris features are considered for identification. Further, an index space based on the dimension of the feature vector is created, which gives a 1-D embedding of the high-dimensional feature set. Next, to minimize the impact of false rejection, the approach considers the possible deviation in each element of the feature vector and then stores the data in possible locations using the predefined threshold. Besides, to reduce the false acceptance rate, linking of the modalities has been done for every individual data. The modality linking thus helps in carrying out an efficient search of the queried data, thereby minimizing the false acceptance and rejection rate. Experiments on a chimeric iris and fingerprint bimodal database resulted in an average of 95% reduction in the search space at a hit rate of 98%. The results suggest that the proposed indexing scheme has the potential to substantially reduce the response time without compromising the accuracy of identification.

Recognition Performance Analysis of a Multimodal Biometric System Based on the Fusion of 3D Ultrasound Hand-Geometry and Palmprint

Multimodal biometric systems are often used in a wide variety of applications where high security is required. Such systems show several merits in terms of universality and recognition rate compared to unimodal systems. Among several acquisition technologies, ultrasound bears great potential in high secure access applications because it allows the acquisition of 3D information about the human body and is able to verify liveness of the sample. In this work, recognition performances of a multimodal system obtained by fusing palmprint and hand-geometry 3D features, which are extracted from the same collected volumetric image, are extensively evaluated. Several fusion techniques based on the weighted score sum rule and on a wide variety of possible combinations of palmprint and hand geometry scores are experimented with. Recognition performances of the various methods are evaluated and compared through verification and identification experiments carried out on a homemade database employed in previous works. Verification results demonstrated that the fusion, in most cases, produces a noticeable improvement compared to unimodal systems: an EER value of 0.06% is achieved in at least five cases against values of 1.18% and 0.63% obtained in the best case for unimodal palmprint and hand geometry, respectively. The analysis also revealed that the best fusion results do not include any combination between the best scores of unimodal characteristics. Identification experiments, carried out for the methods that provided the best verification results, consistently demonstrated an identification rate of 100%, against 98% and 91% obtained in the best case for unimodal palmprint and hand geometry, respectively.

Secure personal authentication in fog devices via multimodal rank‐level fusion

SummaryIn recent years, Fog and edge computing have gained enormous increases and usage of Internet of Things (IoT) devices in research and development even though it provides challenges such as security and privacy. One way to ensure this is a fusion of multimodal biometrics authentication systems, which offer highly reliable biometric authentication solutions to reduce the failure to enroll rate, increase the degree of freedom, and deter spoof attacks. Since unimodal biometric systems often face significant limitations due to sensitivity to noise, data quality, non‐universality, and illumination variations. In this research, we propose management of access control to ensure the desired level of security, a Rank level fusion integration method. ORL, GT, CASIA, and VGG‐face databases were used for testing and evaluating the implemented system. Thus, using the LDA with SCE based feature selection approach, the average GAR percent for the facial, hearing, and palm vein is 3.25% for the facial and muffs pulse and 8.55% for the head and palm vein result was the overall performance of the biometric system even in the presence of low‐quality data in the Fog environment.

Multimodal biometrics system with face, ear, and foot fusion techniques

Biometrics, as an identification method, is used for various applications, particularly in security technologies. The integration of multiple biometric sources aims to overcome limitations observed in unimodal systems, enhancing recognition accuracy. Fusion techniques, categorized into sensor level, feature level, matching score level, decision level, and rank level, are explored to optimize the combination of information from different modalities. Various fusion schemes, such as feature-level fusion, decision-level fusion, and hybrid systems, are investigated for their effectiveness in integrating diverse biometric traits. This paper details the fusion schemes at different levels, including sensor, feature extraction, matching score, and decision levels. Experimental results demonstrate the efficacy of the proposed multimodal biometric system. The Equal Error Rate (EER) is analyzed to evaluate system accuracy, with weights assigned to each modality based on their performance. Normalization techniques and fusion rules are applied to combine modalities, resulting in enhanced matching scores. The analysis of results showcases the performance of the system across various fusion combinations. Notably, the combination of ear and foot modalities yields the highest matching score, demonstrating the effectiveness of fusion techniques in multimodal biometrics.

A Single Objective GA and PSO for the Multimodal Palmprint Recognition System

Biometric plays a vital role in human authentication systems. Unimodal and multimodal biometrics have been active research areas for the past few decades. The investigation of palmprint recognition under various illuminations, rotations, and translations is a challenging task. The research work on multimodal palmprint recognition systems has widely increased to improve the recognition rate and reduce execution time. In this article, a multimodal palmprint biometric system is formed by combining the left and right palmprint images to obtain an optimal recognition rate. A modified multilobe ordinal filter (MMLOF) is used to extract the features. Feature‐level fusion is used to fuse the left and right palmprint images. This results in a high‐dimension feature vector that requires larger memory to store. It creates redundant and irrelevant features that affect the recognition rate. To overcome these limitations, the optimal MMOF features are extracted by optimization techniques such as particle swarm optimization (PSO) and the genetic algorithm (GA). Finally, PSO and GA optimization algorithms are wrapped with the nearest neighbor classifier (NN) to evaluate the fitness function. The experimental analyses are conducted to identify the performance of GA and PSO using the IITD palmprint dataset. The 1st order MMLOF with GA (multimodal) converges faster and outperforms the 1st order MMLOF with PSO (multimodal) and obtains an optimal recognition rate of 96.95%.

NaijaFaceVoice: A Large-Scale Deep Learning Model and Database of Nigerian Faces and Voices

The fusion of two or more traits in multimodal biometrics generally improves recognition accuracy. The question is, by how much? Large-scale databases are better suited for training deep learning models for better generalization and accuracy. Therefore, a large-scale multimodal database is beneficial. However, publicly available large-scale multimodal databases are scarce, especially for faces and voices. Again, because a face image is 2-D while a voice is 1-D, there is the challenge of the best way to fuse both. Therefore, improvements owing to fusion have hitherto yielded marginal improvements. This study proposes a semi-automated curation algorithm for the extraction of the faces and voices of target individuals in videos to create a large-scale face-voice database. The curation technique involves observing the positions at the time of the occurrence of the target subject’s faces and voices in videos. These positions are supplied to a MATLAB2017b script that detects the faces in the observed regions, crops, resizes, auto-labels, and writes them to the disk. A second MATLAB2017b script, extracts the audio content within the observed regions, auto-labels, and writes the voice segments to the disk. The created database named NaijaFaceVoice consists of 2,656 subjects with over 2 million faces and 195 hours of utterances. The database was employed to develop a large-scale recognition system that leveraged Convolutional Neural Networks. Robust fusion methods incorporating the proposed Spectrogram-Voting concept significantly improved performance achieving a record equal error rate of 0.0003519%, an improvement by a factor of over 450.

Fused Multimodal Biometric Authentication for an Open Distance Learning Environment Online Examination

Fused Multimodal Biometric Authentication for an Open Distance Learning Environment Online Examination

Cancellable Multi-Biometric Feature Veins Template Generation Based on SHA-3 Hashing

In this paper, a novel cancellable biometrics technique called Multi-Biometric-Feature-Hashing (MBFH) is proposed. The MBFH strategy is utilized to actualize a single direction (non-invertibility) biometric shape. MBFH is a typical model security conspire that is distinguished in the utilization of this protection insurance framework in numerous sorts of biometric feature strategies (retina, palm print, Hand Dorsum, fingerprint). A more robust and accurate multilingual biological structure in expressing human loneliness requires a different format to record clients with inseparable comparisons from individual biographical sources. This may raise worries about their utilization and security when these spread out designs are subverted as everybody is acknowledged for another biometric attribute.The proposed structure comprises of four sections: input multi-biometric acquisition, feature extraction, Multi-Exposure Fusion (MEF) and secure hashing calculation (SHA-3). Multimodal biometrics systems that are more powerful and precise in human-unmistakable evidence require various configurations to store a comparative customer that can be contrasted with biometric wellsprings of people. Disparate top words, biometrics graphs can’t be denied and change to another request for positive Identifications (IDs) while settling. Cancellable biometrics is may be the special procedure used to recognize this issue.

Novel Multimodal Biometric Feature Extraction for Precise Human Identification

In recent years, biometric sensors are applicable for identifying important individual information and accessing the control using various identifiers by including the characteristics like a fingerprint, palm print, iris recognition, and so on. However, the precise identification of human features is still physically challenging in humans during their lifetime resulting in a variance in their appearance or features. In response to these challenges, a novel Multimodal Biometric Feature Extraction (MBFE) model is proposed to extract the features from the noisy sensor data using a modified Ranking-based Deep Convolution Neural Network (RDCNN). The proposed MBFE model enables the feature extraction from different biometric images that includes iris, palm print, and lip, where the images are preprocessed initially for further processing. The extracted features are validated after optimal extraction by the RDCNN by splitting the datasets to train the feature extraction model and then testing the model with different sets of input images. The simulation is performed in matlab to test the efficacy of the modal over multi-modal datasets and the simulation result shows that the proposed method achieves increased accuracy, precision, recall, and F1 score than the existing deep learning feature extraction methods. The performance improvement of the MBFE Algorithm technique in terms of accuracy, precision, recall, and F1 score is attained by 0.126%, 0.152%, 0.184%, and 0.38% with existing Back Propagation Neural Network (BPNN), Human Identification Using Wavelet Transform (HIUWT), Segmentation Methodology for Non-cooperative Recognition (SMNR), Daugman Iris Localization Algorithm (DILA) feature extraction techniques respectively.

Identification based on feature fusion of multimodal biometrics and deep learning

Identification based on feature fusion of multimodal biometrics and deep learning

Evaluation of Multimodal Biometric Parameters for Diagnosing Acute Angle Closure Secondary to Lens Subluxation.

To evaluate the clinical characteristics and multimodal biometric parameters from ultrasound biomicroscopy (UBM) and IOL Master biometry of patients with acute secondary angle-closure due to lens subluxation (ASAC-LS), acute primary angle-closure (APAC), and cataract. This retrospective study included 22 eyes with ASAC-LS, 27 eyes with APAC, and 39 eyes with cataract. Gender, age, affected eye, best corrected visual acuity, axial length, central corneal thickness, and anterior chamber depth (ACD) assessed by UBM and IOL Master were measured and compared between the three groups. In addition, we compared the ratio of ACD (ACD ratio) and the difference of ACD (ACD difference) measured by the two instruments. Logistic regression analysis was conducted to evaluate the predictive factors for lens subluxation. Receiver operating characteristic (ROC) curves were plotted to obtain a suitable cutoff value of biometric parameters to separate ASAC-LS cases from APAC and cataract cases. In the ASAC-LS group, the median (interquartile range [IQR]) ACD measured by IOL Master was 2.47 (IQR 1.85-2.92) mm while the median ACD measured by UBM was 3.11 (IQR 2.60-3.76) mm. The difference of ACD measured by the two instruments was statistically significant in the ASAC-LS group (P < 0.001) whereas the differences were not statistically significant in the APAC group (P = 0.521) and cataract group (P = 0.204). Subsequent pairwise comparison revealed that only the ACD difference (0.40 [IQR 0.22-1.08] mm) and ACD ratio (1.18 [IQR 1.07-1.40]) in the ASAC-LS group were significantly different from those in the APAC group (ACD difference 0.02 [IQR 0.01-0.07] mm; ACD ratio 1.01 [IQR 1.00-1.04]) and cataract group (ACD difference 0.09 [IQR 0.01-0.14] mm; ACD ratio 1.03 [IQR 1.00-1.04]) (all P < 0.001). The ACD difference and ACD ratio were significantly associated with lens subluxation in the multivariate logistic regression analysis (P < 0.001 and P = 0.001, respectively). Additionally, the ROC curve analysis showed that the ACD difference at 0.235mm and the ACD ratio at 1.080 were the respective cut-off values for lens subluxation, with a sensitivity of 77.3% and specificity of 100.0%. Our findings provide a new option for identifying lens subluxation. Specifically, combining the ACD from UBM and IOL Master may be helpful for differential diagnosis of ASAC-LS.

Modified Firefly Optimization with Deep Learning based Multimodal Biometric Verification Model

Biometric security has become a main concern in the data security field. Over the years, initiatives in the biometrics field had an increasing growth rate. The multimodal biometric method with greater recognition and precision rate for smart cities remains to be a challenge. By comparison, made with the single biometric recognition, we considered the multimodal biometric recognition related to finger vein and fingerprint since it has high security, accurate recognition, and convenient sample collection. This article presents a Modified Firefly Optimization with Deep Learning based Multimodal Biometric Verification (MFFODL-MBV) model. The presented MFFODL-MBV technique performs biometric verification using multiple biometrics such as fingerprint, DNA, and microarray. In the presented MFFODL-MBV technique, EfficientNet model is employed for feature extraction. For biometric recognition, MFFO algorithm with long short-term memory (LSTM) model is applied with MFFO algorithm as hyperparameter optimizer. To ensure the improved outcomes of the MFFODL-MBV approach, a widespread experimental analysis was performed. The wide-ranging experimental analysis reported improvements in the MFFODL-MBV technique over other models.

Power Minimization by Designing Novel Architecture for Multimodal Biometric System

Abstract: The reduction of power, area, and delay parameters are increasing as the range of sophistication of applications in VLSI circuit designs. Adders and multipliers play a vital role in VLSI circuit designs to minimize the power consumed by the circuit. The Finite Impulse Response Filter is widely used in Digital Signal Processing Applications, such as speech processing, loudspeaker equalization, echo cancellation, noise cancellation, arithmetic computations and, image-processing applications. In FIR filter, the adders and multipliers are the important components to reduce the power, delay and area. Therefore, this proposed method faced with more constraints: as minimal power as possible. This work proposes a design of low power Finite Impulse Response Filter using Carry Skip Adder.The recent electronics industry uses digital filters for various real time applications. A novel VLSI architecture is designed by using ALU based FIR filters for biometric system. Where in ALU includes adders and multipliers, comparing different adders for there better performance with respect to power and identifying carry skip adder will consume less power compared to other adders. power consumption of overall architecture by including carry skip adder is very less and we can use this same adder for multimodal architecture also.