A  Survey  of  Biometric  Recognition  Using  Deep  Learning

Mehraj, Haider; Mir, Ajaz Hussain

doi:10.4108/eai.27-10-2020.166775

Cited by 4 publications

(5 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Fingerprints have widely been used in different domains including criminal investigations and in digital security systems to facilitate criminal investigations and strengthen security systems, respectively. There are researches conducted on the application of DL in biometric recognition and authentication schemes have been surveyed in [3,[27][28][29][30][31][32]. However, these previous studies mainly focused on a particular application area, such as security [29], fingerprint liveness detection [28], fingerprint classification [30], fingerprint enhancement [32] etc.…”

Section: Open Accessmentioning

confidence: 99%

“…However, these previous studies mainly focused on a particular application area, such as security [29], fingerprint liveness detection [28], fingerprint classification [30], fingerprint enhancement [32] etc. or study based on biometrics with fingerprint inclusive [3,27,31] and some are not DL based surveys on the contrary [28,29,32]. In addition, those surveys were not systematic and did not present the applications of DL in fingerprint biometrics analyses based on different areas of applications.…”

Section: Open Accessmentioning

confidence: 99%

See 1 more Smart Citation

Deep Learning Algorithms based Fingerprint Authentication: Systematic Literature Review

Chiroma¹

2021

AIS

View full text Add to dashboard Cite

Deep Learning algorithms (DL) have been applied in different domains such as computer vision, image detection, robotics and speech processing, in most cases, DL demonstrated better performance than the conventional machine learning algorithms (shallow algorithms). The artificial intelligence research community has leveraged the robustness of the DL because of their ability to process large data size and handle variations in biometric data such as aging or expression problem. Particularly, DL research in automatic fingerprint recognition system (AFRS) is gaining momentum starting from the last decade in the area of fingerprint pre-processing, fingerprints quality enhancement, fingerprint feature extraction, security of fingerprint and performance improvement of AFRS. However, there are limited studies that address the application of DL to model fingerprint biometric for different tasks in the fingerprint recognition process. To bridge this gap, this paper presents a systematic literature review and an insightful meta-data analysis of a decade applications of DL in AFRS. Discussion on proposed model’s tasks, state of the art study, dataset, and training architecture are presented. The Convolutional Neural Networks models were the most saturated models in developing fingerprint biometrics authentication. The study revealed different roles of the DL in training architecture of the models: feature extractor, classifier and end-to-end learning. The review highlights open research challenges and present new perspective for solving the challenges in the future. The author believed that this paper will guide researchers in propose novel fingerprint authentication scheme.

show abstract

Section: Open Accessmentioning

confidence: 99%

Section: Open Accessmentioning

confidence: 99%

Deep Learning Algorithms based Fingerprint Authentication: Systematic Literature Review

Chiroma¹

2021

AIS

View full text Add to dashboard Cite

show abstract

“…Following the same reasoning described in Section 3.1, we opted for using the normalized square Euclidean distance defined in Equation (3) as the pairwise vectors' comparison measure. Additionally, the definition of the neighborhood is again computed only on the instructor's side, similar to the definition provided in Equation (5). To construct the loss function, we first define a measure of Local Affinity Contrast for a set of N feature vectors with neighborhood mask M and normalized pairwise distances D, as follows:…”

Section: Affinity Contrast Lossmentioning

confidence: 99%

“…For a very large number of important applications, the necessity of acquiring a large body of training data to benefit from learning features tailored to the task-at-hand renders the end-to-end learning of deep models prohibitive. Data acquisition and annotation in several fields, such as biometric recognition, forensics, biomedical imaging, etc., is notoriously difficult due to various restrictions and limitations [5][6][7] (e.g., cost of specialized personnel, privacy issues, etc.). Therefore, research and development in such fields can greatly benefit [8] from techniques that enable powerful algorithms such as CNNs to efficiently learn from limited datasets, or equivalently increase the performance of the current techniques for training under such restrictions.…”

Section: Introductionmentioning

confidence: 99%

Geometric Regularization of Local Activations for Knowledge Transfer in Convolutional Neural Networks

2021

View full text Add to dashboard Cite

In this work, we propose a mechanism for knowledge transfer between Convolutional Neural Networks via the geometric regularization of local features produced by the activations of convolutional layers. We formulate appropriate loss functions, driving a “student” model to adapt such that its local features exhibit similar geometrical characteristics to those of an “instructor” model, at corresponding layers. The investigated functions, inspired by manifold-to-manifold distance measures, are designed to compare the neighboring information inside the feature space of the involved activations without any restrictions in the features’ dimensionality, thus enabling knowledge transfer between different architectures. Experimental evidence demonstrates that the proposed technique is effective in different settings, including knowledge-transfer to smaller models, transfer between different deep architectures and harnessing knowledge from external data, producing models with increased accuracy compared to a typical training. Furthermore, results indicate that the presented method can work synergistically with methods such as knowledge distillation, further increasing the accuracy of the trained models. Finally, experiments on training with limited data show that a combined regularization scheme can achieve the same generalization as a non-regularized training with 50% of the data in the CIFAR-10 classification task.

show abstract

“…For an authentication system to instill total confidence that a person actually is who they say they are, there are three ways, depending on what they know, what they have, or who it is [16]- [18]. For this project, "who it is" was used with the help of biometrics in the area of face and voice recognition, which are unique personal characteristics and parameters [19], [20]. In order to strengthen the authentication system, Deep Learning techniques were applied [21], [22] in the fields of artificial vision, natural language processing, and speech processing [19].…”

Section: Introductionmentioning

confidence: 99%

Access Control System based on Voice and Facial Recognition Using Artificial Intelligence

Capote-Leiva¹,

Villota-Rivillas²,

Muñoz³

2022

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

Computer security has become a matter of great concern at the global level. Whatever the economic sector, all companies handle confidential information related to clients and personnel. These latter therefore need to be seen as sensitive assets that require protection. Appropriate, consensual handling of personal information is thus a legal and delicate requirement that demands to be treated securely in all types of business. Many companies and clients have lost sums of money in the millions due to incorrect information protection, triggering complicated, expensive proceedings that are awkward and cumbersome to resolve. To implement an authentication system based on biometric parameters, which strengthens the security of sensitive assets in areas considered critical in various organizations by attaining the highest accuracy in user classification processes. By applying a convolutional neural network: MobileNet, viewing via computer and low-cost devices (Raspberry Pi 3). The system constitutes an authentication device for voice and face with 96% and 100% accuracy, respectively. Conclusion: The system shows that deep learning (Deep Convolutional Neural Networks), in combination with devices such as the Raspberry, generate a system capable of high performance in time and cost, as well as providing companies with a robust system featuring high accuracy in the correct recognition of the biometric patterns of users registered and trained by the system.

show abstract

A Survey of Biometric Recognition Using Deep Learning

Cited by 4 publications

References 13 publications

Deep Learning Algorithms based Fingerprint Authentication: Systematic Literature Review

Deep Learning Algorithms based Fingerprint Authentication: Systematic Literature Review

Geometric Regularization of Local Activations for Knowledge Transfer in Convolutional Neural Networks

Access Control System based on Voice and Facial Recognition Using Artificial Intelligence

Contact Info

Product

Resources

About