Fingerprint Recognition Algorithm Research Articles

Robust fingerprint recognition approach based on diagonal slice of polyspectra in the polar space

Although fingerprint recognition is a mature technology and nowadays commercial state-of-the-art systems can be successfully used in a number of real applications, not all the problems have been solved and the research is still very active in the field. This paper presents a new approach for estimating the shift and rotation parameters between fingerprint images stored in the dadabase that operates in the third-order frequency-domain measure called the auto-bispectrum, which, allows us to estimate the shift and the rotation separately. The diagonal slices and their spectra of auto- and cross-bispectrum are proposed. The rotation parameters are estimated from the remaining polar sampled the third-order spectrum using cross-correlation, and then, after compensating for rotation, we may easily estimate the translational component, e.g., by using phase correlation.Experimental evidence of this performance is presented, and the mathematical reasons behind these characteristics are explained in depth. We compare our approach in a simulation to other frequency-domain fingerprint recognition algorithms. We find that our algorithm can better estimate shift and rotation parameters than the other methods.

Developing an algorithm for Fingerprint Recognition of Newborns and Toddlers based on Deep Neural Networks and Transfer Learning

In the late 19th century, Sir Galton started the use of fingerprint recognition as a method for baby identification. However, even after the extent two centuries, fingerprint identification for newborns and toddlers has not advanced to the same level as that for adults. According to research from the International Centre for Missing and Exploited Children, there is a growing request for fingerprint identification of babies because more than a million of them go missing each year. So, we want to ensure that newborns and toddlers have access to their rights, including vaccinations, healthcare, and nutritional supplements until they reach school age. This paper develops an algorithm that has been used in previous research to recognize newborns and toddlers to produce remarkable and promising results by Pre-trained model deep neural networks and transfer learning to extract the deep features and to get the identification score, aiming to enhance accuracy and validate. The proposed model is validated on the fingerprint dataset of babies named as NITG dataset. The fingerprint recognition model proposed in the research showcases excellent results, where achieved 100% accuracy in training and 90.35% in validation.
  

Radio Frequency Fingerprinting Identification of Few-Shot Wireless Signals Based on Deep Metric Learning

As a cross-protocol endogenous security mechanism, the physical layer-based radio frequency (RF) fingerprinting can effectively enhance the existing password-based application layer authentication utilizing the hardware differences of wireless devices, which is unique and cannot be counterfeited by a third party. However, the recognition performance of the deep learning physical layer fingerprint recognition algorithm drops sharply in the case of a small number of signal samples. This paper analyzes the feasibility and proposes the few-shot wireless signal classification network based on deep metric learning (FSig-Net). FSig-Net reduces the model’s dependence on big data by adaptively learning the feature distance metric. We use 8 mobile phones and 18 Internet of Things (IoT) modules as targets for identification. When the number of single-type samples is only 10, the recognition accuracy of mobile phones can reach 98.28%, and the recognition accuracy of IoT devices can reach 98.20%.

A novel acoustic emission signal segmentation network for bearing fault fingerprint feature extraction under varying speed conditions

The acoustic emission (AE) technique is known for its sensitivity to early damage and is well-suited for online condition monitoring of rolling element bearings (REBs) in various industrial application scenarios. Nonetheless, identifying weak faults under varying speed and strong background noise conditions still remains challenging. In addition, the comprehensive modelling of AE signal from faulty REB in electromechanical systems is still a pending issue. In light of this, a well-considered model is firstly developed for the AE signal of faulty REBs in this work. After that, a novel bearing fault diagnosis framework based on semantic segmentation networks is devised. Precisely, the proposed framework consists of three main components: a preprocessing step depending on the signal segmentation algorithm, a diagnosis step using fault fingerprint mapping, and a postprocessing evaluation step supplemented by a density peak clustering (DPC) approach. We evaluate the presented procedures through simulation analysis and an experimental case under varying speed conditions. Meanwhile, the comparison with the original threshold-based fault fingerprint recognition algorithm is also conducted. The comprehensive results demonstrate the efficacy of identifying fault-associated fingerprint feature (FPF), indicating that the proposed framework holds promise for condition monitoring.

Biometric based Mobile Voting Application

With a population of more than one billion, India is the largest democracy in the world. Its electorate is more than 668 million people and includes 543 congressional districts. Each Indian citizen above 18 years of age is entitled to vote in elections to choose their leaders. The traditional method of voting involved paper ballot voting and at present Electronic Voting Machines (EVM) are employed. The goal of the mobile voting system research is to create a software that makes voting simple in all elections. This system aims to increase election security by using a fingerprint recognition algorithm for voter identity authentication. This allows all eligible voters to participate in the election from any place. There are some circumstances like COVID epidemic where a physical election is not feasible, and this model overcomes it. It helps to provide reliable, timely, and accurate election results. Voter ID, Aadhar ID and biometric system (fingerprint) are used by the website for user login.

Application of fingerprint image fuzzy edge recognition algorithm in criminal technology

Abstract In the context of the rapid development of science and technology and the modernization of the legal system, criminal activities are becoming more and more intelligent and technological, which also puts forward higher requirements for criminal technology. The current criminal technology equipment is relatively backward, and the technical level is not high enough, resulting in a low utilization rate of trace material evidence extraction, which directly affects the role of criminal technology in the investigation and solving of cases. In recent years, fingerprint recognition algorithms and image edge detection algorithms have been widely used in various fields. This work studied the application of fingerprint image fuzzy edge recognition algorithm in criminal technology, in order to improve the level of criminal technology and the utilization rate of physical evidence extraction. The criminal technology system is upgraded and optimized by combining fingerprint recognition algorithm and image edge detection algorithm. And fuzzy theory is added to ensure the feasibility of the research. The experimental results show that the fuzzy edge recognition algorithm of fingerprint image can improve the level of criminal technology and the utilization rate of material evidence to a certain extent. The utilization rate is increased by 7.04%. The recognition accuracy of the fuzzy recognition method is also 13.2% higher than that of the methods in the literature.

Subjecting soleprint to fingerprint recognition algorithm treatment

This experiment was carried out to demonstrate that a Soleprint would easily substitute a fingerprint for identification and voting purposes for leprosy patients, injury victims, and natural causes who have no fingerprints. The fingerprint recognition algorithm was used throughout the experiment to conduct preprocessing stages such as segmentation, normalization, Gaussian blurring, noise reduction, Canny, Orientation, Binarization Thinning, Gabor Filtering, and Minutiae Extraction. These stages were performed on a Soleprint image and a Fingerprint image with the same properties of 8-bit grey-level, 600-dpi resolution, and scale 256 pixels, all in bitmap image compression format. Since the experiment was not designed to assess recognition algorithms, computation time was not considered. When the findings were compared, it was discovered that the algorithm has the same impact on both photos (Soleprint and Fingerprint, both whorl in class) and that the minutiae points are the basis for matching a fingerprint in today's trend, are also present on the Soleprint. A modern method of self-identity and voting for lepers and injury patients is now accessible.

Effects of fingerprint recognition algorithms on toeprint and fingerprint

Democracy is a form of government in which politicians are chosen by popular vote. Voting may be done using behavioural biometrics, known as transparent balloting, which involves indicating a candidate of preference by indication, or physiological biometrics, known as hidden balloting, which involves appending a fingerprint to a vote. In certain countries around the world, fingerprints are used to cast ballots. However, injury casualties, natural disaster victims, and lepers, who account for one out of every 10,000 people in the world's population of approximately 7.7 billion, may pose a concern if their fingerprints are used to classify them. Two photographs were chosen to make up the sample collection: a Toeprint image with a left loop to be contrasted with a fingerprint image with a left loop. The photos were subjected to the standard fingerprint recognition algorithm, which included processes such as Normalisation, Orientation, Binarization, Thinning, and Minutiae Extraction. Since the experiment was not designed to measure the accuracy of the recognition algorithm, processing time and image quality were not taken into account. When the findings were compared, it was discovered that the algorithm had the same impact on both photos and that the derived minutiae points, which are the foundation for successful matching, were greater than 100. A modern method of self-identity and voting for lepers and injury patients is now accessible.

Security-Level Improvement of IoT-Based Systems Using Biometric Features

The Internet of Things (IoT) is reported as a main research topic in the current decade. It will be possible to connect smart devices to each other using IoT, a platform such as the Internet. However, the expansion and intrusion of such a large network raises some new security issues and risks related to the disclosure of user confidential information where these devices are subject to hacker threats and intrusions. Traditional security systems were password based. In this paper, after reviewing the actions taken in this regard, the improvement level of biometric security compared with traditional password-based methods will be proven in section three using the Markov model. By considering the results of the evaluation, the probability of occurrence of security problems is decreased by 90.71% by applying biometric features. Then, multi-layer security architecture with biometric features and coding systems is suggested to increase security. In the first layer, the fingerprint recognition algorithm is dependent on the module, and the U.are.U 5100 module provides more security than others. In the second layer, the Hash mechanism of the MD5 algorithm is, on average, 63.21% more efficient. By determining the properties of the first two architectural layers and ultimately for the IoT application layer, empirical methods and hardware platforms for the Internet of things are used. Concerning the simulation results, the suggested mechanism enhances the system security by 120.38% on average, which is 106.23, 110.45, and 144.46% of relative improvement compared with IoT sensors, controller layer mechanisms, and application layer mechanisms, respectively.

Effects of fingerprint recognition algorithms on toeprint and fingerprint

Democracy is a form of government in which politicians are chosen by popular vote. Voting may be done using behavioral biometrics, known as transparent balloting, which involves indicating a candidate of preference by indication, or physiological biometrics, known as hidden balloting, which involves appending a fingerprint to a vote. In certain countries around the world, fingerprints are used to cast ballots. However, injury casualties, natural disaster victims, and lepers, who account for one out of every 10,000 people in the world's population of approximately 7.7 billion, may pose a concern if their fingerprints are used to classify them. Two photographs were chosen to make up the sample collection: a Toeprint image with a left loop to be contrasted with a fingerprint image with a left loop. The photos were subjected to the standard fingerprint recognition algorithm, which included processes such as Normalisation, Orientation, Binarization, Thinning, and Minutiae Extraction. Since the experiment was not designed to measure the accuracy of the recognition algorithm, processing time and image quality were not taken into account. When the findings were compared, it was discovered that the algorithm had the same impact on both photos and that the derived minutiae points, which are the foundation for successful matching, were greater than 100. A modern method of self-identity and voting for lepers and injury patients is now accessible.

An optimized recognition algorithm for SSL VPN protocol encrypted traffic

With the widespread use of Virtual Private Networks (VPNs), the identification of Secure Sockets Layer (SSL) VPN encrypted traffic has become an important issue. This paper first introduced SSL VPN encrypted traffic and analyzed the flow of its handshake protocol. Then, an improved fingerprint recognition algorithm was designed to identify SSL streams. Capsule Neural Network (CapsNet), an optimized convolutional neural network, was used to recognize SSL VPN. An experimental analysis was carried out on the ISCXVPN2016 dataset. It was found that the recognition accuracy of the proposed method reached up to 99.98% for SSL streams, and the convergence speed was high; the recognition precision reached 98.16%, and the recall rate reached 99.98% for SSL VPNs, both of which were better than the algorithms such as random forest (RF) and C4.5. The experimental results verify the effectiveness of the optimized recognition algorithm for SSL VPN recognition and make some contributions to its application in practice.

Dactylogram Processing System

Recognition of fingerprints (dactyloscopic images) is one of the practical application of signal processing. System of person identification by fingerprints is commonly-used by law enforcement bodies and Border services. This is also important in the field of access control systems and commercial devices where data security is not less important as reliability and data rate of processing algorithms. Existing systems of fingerprints processing are not fully ready for automatic recognition. Also, full modernization of existing equipment is not possible. The paper is devoted to the method of image processing. In particular, the preliminary processing of dactyloscopic images is considered as well as development of theoretical approach and practical realization of first stage of patterns forming – pre-processing of image for decreasing of its size and contrast increasing. The criteria for selecting ranges for sampling and quantization of images are given. Tasks of reducing the fingerprint image while increasing the contrast of the image were considered, analyzed and solved. Image reduction is based on the use of interpolation. It is shown that among the considered interpolation methods - linear, bilinear and bicubic - the latter one could provide the highest accuracy although it needs more hardware resources. However, when the dpi parameter (dots-per-inch) falls below 150, a rapid increase in the number of artifacts in the image is observed. Increasing of image sharpness is necessary for highlighting of colour transitions and consequently – for increasing the percentage of correct recognitions. Such increasing of image sharpness is proposed to achieve by using the Laplace operator (Laplasian calculation) and adding the result to the original image. The value of derivative at each pixel of the image depends linearly on sharpness level. Thus, it allows separating the areas with abrupt colour changes and gaps from the areas where the brightness is constant or changes slowly. The result of second derivative is much more for the areas with sharp changes than for the areas without them. The areas with constant or slowly-changing brightness after the second derivative calculation become almost the same dark colour. These areas could be restored to original image with retention of sharpness increasing effect. For this, transformed by Laplasian image should be added to the original one. Use of Laplasian allows to get an acceptable balance between the speed and computational complexity of the fingerprint recognition algorithm. The technical implementation of the device and illustration of its operation are given. Fingerprints image processing system is executed on the base of STM32f407 microcontroller with CortexM core. The system includes capasitive scanner, TFT LCD display and lab power source. The microcontroller software realizes, in particular, interpolation and contrast increasing. The system is module-compatible and able for scaling.

A multi-task fully deep convolutional neural network for contactless fingerprint minutiae extraction

With the outbreak and wide spread of novel coronavirus (COVID-19), contactless fingerprint recognition has attracted more attention for personal recognition because it can provide significantly higher user convenience and hygiene than the traditional contact-based fingerprint recognition. However, it is still challenging to achieve a highly accurate recognition due to the low ridge-valley contrast and pose variances of contactless fingerprints. Minutiae points are a kind of ridge flow discontinuities, and robust and accurate extraction is an important step for most automatic fingerprint recognition algorithms. Most of existing methods are based on two stages which locate the minutiae points first and then compute their directions. The two-stage method cannot make full use of location and direction information. In this paper, we propose a multi-task fully deep convolutional neural network for jointly learning the minutiae location detection and its corresponding direction computation which operates directly on the whole gray scale contactless fingerprints. The proposed method consists of offline training and online testing stages. In the training stage, a fully deep convolutional neural network is built for the tasks of minutiae detection and its direction regression, with an attention mechanism to make the direction regression branch concentrate on the minutiae points. A new loss function is proposed to jointly learn the tasks of minutiae detection and its direction regression from the whole fingerprints. In the testing stage, the trained network is applied on the whole contactless fingerprint to generate the minutiae location and direction maps. The proposed multi-task leaning method performs better than the individual single task and it operates directly on the raw gray-scale contactless fingerprints without preprocessing. The results on three contactless fingerprint datasets show the proposed algorithm performs better than other minutiae extraction algorithms and the commercial software.

Application of Machine Learning in Fingerprint Image Enhancement and Recognition: A Review

Biometric characteristics helps to recognize an individual among others. Each individual has a unique biometric feature. So, an automated system is designed to recognize an individual. In today’s growing AI development, biometric recognition is applied in many security systems. One of oldest and widely used authentic biometric methodology is fingerprint recognition. Many fingerprint recognition algorithms are designed and developed in order to reduce error rate and to improve accuracy. In this paper, a comprehensive review is presented on various techniques used for fingerprint recognition system along with their performance and their limitations. The purpose of this paper is to review various recent work on the fingerprint recognition system, to explain step by step the steps for recognizing fingerprints, and to provide summaries of the fingerprint databases with functionality

Feature extraction, recognition, and matching of damaged fingerprint: Application of deep learning network

SummaryWith the improvement of informatization and the development of computer technology, identity recognition with fingerprint has been the growing trend, but the stained fingerprint will make recognition difficult. To solve the above problem, this paper briefly introduced the traditional point matching fingerprint recognition algorithm and the damaged fingerprint recognition method based on the convolution neural network (CNN). Then in MATLAB software, the damaged fingerprint recognition method based on CNN was simulated and compared with the traditional point matching recognition method and traditional CNN recognition method. The results showed that the improved CNN recognition method iterated fewer times and achieved smaller training error during the training; the improved CNN method had lower false recognition rate and rejection rate in identifying the unknown fingerprint; the traditional point matching method spent the most time in identifying the unknown fingerprint, followed by the traditional CNN method and the improved CNN recognition method. This study used CNN for fingerprint recognition and improved CNN to improve its recognition accuracy for the damaged fingerprint. The improved CNN can effectively enhance the recognition accuracy of the damaged fingerprint, which provides a useful reference for the improvement of the accuracy and applicability of the fingerprint recognition system.

A Flexible Touch-Based Fingerprint Acquisition Device and a Benchmark Database Using Optical Coherence Tomography

Optical coherence tomography (OCT) is a high-resolution imaging technology probing the internal structure of multilayered tissues. Since it provides subsurface fingerprint information that is identical to the surface texture but unaffected by any surface defects, OCT-based fingerprints open up a new domain for establishing robust and high-security automatic fingerprint identification systems (AFISs). However, the development of OCT-based fingerprint recognition is hindered by the lack of public OCT-based fingerprint database for algorithm analysis and testing. This article, for the first time, established an OCT-based fingerprint database with thousands of fingers using our custom-built acquisition device. The website of this data set is https://github.com/CV-SZU/. Moreover, the images included in the database were selected after quality evaluation based on image resolution, image size, effective measured area, and the number of extractable features. Finally, case studies, including antispoofing, multiple subsurface fingerprint reconstruction, and fingerprint verification, were discussed based on the developed database. The database can serve as a benchmark for developing effective antispoofing, live detection, and high-accurate fingerprint recognition algorithms. It will significantly promote the research in the area of fingerprint biometric and will also advance the development of commercial products.

Cancelable face and fingerprint recognition based on the 3D jigsaw transform and optical encryption

Biometric systems are widely used now for security applications. Two major problems are encountered in biometric systems: the security problem and the dependence on a single biometric for verification. The security problem arises from the utilization of the original biometrics in databases. So, if these databases are attacked, the biometrics are lost forever. Hence, there is a need to secure original biometrics by keeping them away from utilization in biometric databases. Cancelable biometrics is an emerging security trend in the field of biometric authentication. Cancelable biometric systems depend on the transformation of biometric features into new formats so that users can replace their biometric templates in the same or different systems. In this paper, we present a proposed cancelable face and fingerprint recognition algorithm based on the 3D jigsaw transform and optical encryption. The algorithm adopts the Fractional Fourier Transform (FRFT) in the optical encryption scheme with a single random phase mask. This structure can be implemented all optically with a single lens. The proposed cancelable biometric recognition algorithm employs an optical image encryption scheme that depends on two cascaded stages of 2D-FRFT with separable kernels in both dimensions. The two stages are separated with a random phase mask. A preceding bit plane permutation process is performed on the obtained biometrics prior to the FRFT operation to achieve a high level of security. To validate the proposed algorithm for cancelable biometric recognition, different sets of face and fingerprint images are used. A comparative study is presented between the proposed algorithm and the optical Double Random Phase Encoding (DRPE) algorithm. The simulations results obtained for performance evaluation show that the proposed algorithm is safe, reliable, and feasible. It has good encryption and cancelability that reveal good performance.

Modelling and Simulation of Robust Biometric Fingerprint Recognition Algorithm

Fingerprint is a type of biometrics that is used for human recognition because it is rich in details called minutiae. The focus of this paper is to develop an algorithm for fingerprint recognition through extracting and matching of fingerprint image minutiae. In order to achieve accurate results, the fingerprint is first pre-processed to enhance its quality. An alignment-based matching algorithm was developed for minutia matching. This algorithm is capable of finding the correspondences between the input minutia pattern and the stored template minutia pattern without resorting to exhaustive search. The performance of the developed system was evaluated with matrix laboratory (MATLAB) software along with fingerprint from the standard FVC2000 fingerprint database and fingerprints from different people at university of Uyo. From the curves of false match rate (FMR) and false non match rate (FNMR) against similarity score, a threshold value of 0.4 was obtained as guided by the equal error rate (EER) value. For verification and identification testing, the developed system was able to accurately differentiate between genuine users and imposter irrespective of the orientation and noise level of the fingerprints.

Inkjet-Printed Carbon Nanotubes for Fabricating a Spoof Fingerprint on Paper.

A spoof fingerprint was fabricated on paper and applied for a spoofing attack to unlock a smartphone on which a capacitive array of sensors had been embedded with a fingerprint recognition algorithm. Using an inkjet printer with an ink made of carbon nanotubes (CNTs), we printed a spoof fingerprint having an electrical and geometric pattern of ridges and furrows comparable to that of the real fingerprint. With this printed spoof fingerprint, we were able to unlock a smartphone successfully; this was due to the good quality of the printed CNT material, which provided electrical conductivities and structural patterns similar to those of the real fingerprint. This result confirms that inkjet-printing CNTs to fabricate a spoof fingerprint on paper is an easy, simple spoofing route from the real fingerprint and suggests a new method for outputting the physical ridges and furrows on a two-dimensional plane.

RF Fingerprint Feature Extraction and Recognition Algorithm Based on Double-Maximal

The proposed double-maximal radio frequency (RF) fingerprint feature extraction algorithm in this paper is aimed at the imperfect signal with noisy signal. It can achieve perfect recognition rate when the feature dimension is greatly reduced, and the classification performance is improved. The advantage of low computational complexity and high authentication accuracy is suitable for access authentication in scenarios such as power terminals, and is well compatible with the diversified and heterogeneous characteristics of power terminals under the power edge computing.