Iris Data Research Articles

A New Feature Extraction Approach in Classification for Improving the Accuracy in Iris Recognition

Personal identity is becoming increasingly vital to meet the increasing security standards of today's business society. Iris recognition is one of the most accurate biometric technologies currently in use. Iris recognition is employed in high-security sectors due to its dependability and flawless identification rates. The steps of iris identification, comprising image preparation, extraction of features, and classifier creation, are described thoroughly in the primary portion of this research. The feature extraction stage is the most important in an iris identification system since it extracts the iris's distinctive feature. Several methods have been devised to extract the various characteristics that are unique to everyone. Modern iris identification systems frequently use Gabor filters to identify iris textural characteristics. However, in the application, it is necessary to identify the appropriate Gabor modules and to generate a pattern of iris Gabor characteristics. This research aims to provide a novel multi-channel Gabor filter and Wavelet filter for breaking down and extracting iris data from two different iris datasets. Because wavelet is the most scalable method of image processing, the research investigates using it to create a unique pattern for the iris recognition system. The MATLAB program is used to implement these ideas. CASIA and MMU are the datasets used for this purpose, and their comparative analysis is addressed in the research. To show how well the method performs, experimental results are given. We demonstrate through experiments that the suggested approach results in excellent iris identification performance.Â

Iris Data Compression Based on Hexa-Data Coding

Iris research is focused on developing techniques for identifying and locating relevant biometric features, accurate segmentation and efficient computation while lending themselves to compression methods. Most iris segmentation methods are based on complex modelling of traits and characteristics which, in turn, reduce the effectiveness of the system being used as a real time system. This paper introduces a novel parameterized technique for iris segmentation. The method is based on a number of steps starting from converting grayscale eye image to a bit plane representation, selection of the most significant bit planes followed by a parameterization of the iris location resulting in an accurate segmentation of the iris from the original image. A lossless Hexadata encoding method is then applied to the data, which is based on reducing each set of six data items to a single encoded value. The tested results achieved acceptable saving bytes performance for the 21 iris square images of sizes 256x256 pixels which is about 22.4 KB on average with 0.79 sec decompression average time, with high saving bytes performance for 2 iris non-square images of sizes 640x480/2048x1536 that reached 76KB/2.2 sec, 1630 KB/4.71 sec respectively, Finally, the proposed promising techniques standard lossless JPEG2000 compression techniques with reduction about 1.2 and more in KB saving that implicitly demonstrating the power and efficiency of the suggested lossless biometric techniques.

Comparison of Swarm-based Metaheuristic and Gradient Descent-based Algorithms in Artificial Neural Network Training

This paper aims to compare the gradient descent-based algorithms under classical training model and swarm-based metaheuristic algorithms in feed forward backpropagation artificial neural network training. Batch weight and bias rule, Bayesian regularization, cyclical weight and bias rule and Levenberg-Marquardt algorithms are used as the classical gradient descent-based algorithms. In terms of the swarm-based metaheuristic algorithms, hunger games search, gray wolf optimizer, Archimedes optimization, and the Aquila optimizer are adopted. The Iris data set is used in this paper for the training. Mean square error, mean absolute error and determination coefficient are used as statistical measurement techniques to determine the effect of the network architecture and the adopted training algorithm. The metaheuristic algorithms are shown to have superior capability over the gradient descent-based algorithms in terms of artificial neural network training. In addition to their success in error rates, the classification capabilities of the metaheuristic algorithms are also observed to be in the range of 94%-97%. The hunger games search algorithm is also observed for its specific advantages amongst the metaheuristic algorithms as it maintains good performance in terms of classification ability and other statistical measurements.

The discrete wavelet transform based iris recognition for eyes with non-cosmetic contact lens

Iris recognition has been used as one of the biometric systems for user authentication, identification, and verification for quite some time. The basis of an iris recognition lies on the matching algorithm, which requires similarities of the iris data in the database with the captured one. In addition, nowadays using non-cosmetic or prescribed contact lenses becomes more popular and more preferred choice of many people, which makes the number of contact lens wearers significantly increases. These eyes with contact lenses add more complexity to the iris recognition process, since it can disturb the matching process which then affect the performance of the system. This situation has motivated this study to propose an iris recognition system that works for eyes with lenses. The proposed iris recognition system for eyes wearing non-cosmetic contact lenses was developed using circular hough transform (CHT) in the preprocessing phase for iris localization and discrete wavelet transform (DWT) for feature extraction. Experiments conducted on the proposed system has shown promising results with good accuracy of 0.95 for eyes with no contact lens and 0.8 for eyes with non-cosmetic lenses. The findings also suggested the importance of the iris localization process to the performance of the recognition system.

Design of Time-Encoded Spiking Neural Networks in 7-nm CMOS Technology

In biologically inspired spiking neural networks (SNNs) neurons communicate by short pulses, called spikes. SNNs have the potential to be more power efficient than artificial neural networks (ANNs), thanks to the fewer computational steps required by the spike transmission and processing, as compared to the multiply-and-accumulate (MAC) operations with wide bit-vectors usually adopted in ANNs. We present the design of two types of SNNs with integrate-and-fire dynamics and single-spike per neuron operation, where neural communication is based on synchronous time-to-first-spike (sTTFS) and time-to-first-spike (TTFS) encoding schemes. In the considered time-encoded SNNs, the information is carried by the timing of the spikes with respect to a reference time. In 7nm CMOS technology both designs are synthesized as VHDL-based random-logic-macros (RLMs) and compared to an equivalent ANN design in terms of power consumption, latency and silicon area, using the Iris data set for inference. A cost function expressed as a product of energy consumption and silicon area is introduced to compare the three network designs. With respect to this cost function, it turns out that the SNN-TTFS implemented for the considered classification task outperforms the ANN used as baseline model.

A convolutional neural network based classification for fuzzy datasets using 2-D transformation

Researches on deep learning methods have been actively conducted for the past 10 years, and various deep learning techniques have been proposed by many researchers. In addition, prediction methods using deep learning are widely used in various fields. In particular, convolution neural network (CNN) is most commonly applied to analyze visual images, but it can be also applied to many other data. On the other hand, fuzzy theory has been applied to deep learning techniques in traffic problem, agriculture, and airline customer service. In the case of data containing ambiguous information, data analysis can be performed using soft methods. In particular, the fuzzy theory is widely used to deal with such data. So, when the data includes vague information a fuzzy number can be applied to input/output data. In this paper, seven models using CNN have been proposed to analyze fuzzy input containing ambiguous or linguistic information. Our proposed models use five activation functions. For the data analysis, three datasets including Iris data, US Health Insurance data, Wine quality data are used to compare the seven proposed Fuzzy CNN models.

Modeling Crustal Structure in West Sumatera Using Teleseismic Receiver Function

Indonesia is located between three significant plate confluences that play a role in creating a subduction zone. Sumatra, one of Indonesia’s provinces, is located near the collision of the Eurasian Plate, the Indo-Australian Plate, and the Pacific Plate. The Semangko Fault, the Mentawai Islands, a chain of non-volcanic forearc islands, and active faults that induce earthquakes and landslides are all produced by subduction zones on Sumatra Island. The receiver function approach was developed to identify the discontinuity limits of the Earth’s structures, such as the crust and mantle, from teleseismic events. The information in this study comes from 25 temporal and 1 permanent station in West Sumatra and the time period between 2007 and 2009 from the IRIS data station. The earthquake data was taken from teleseismic earthquakes from the three components of seismic stations using criteria arc distances between 30° and 90° and magnitudes ranging from 6.0 to 10.0. This research aims to determine how crustal thickness is distributed and what effect regional geological structure has on crustal thickness. According to the early results of the study, the estimated range of crustal thickness in the research region is 27.11–32.48 km, and the range of Vp/Vs ratio values is 1.70–2.19. The direction of crustal thickness thickening in the area was likewise East-South.

Two-Factor Biometric Identity Verification System for the Human-Machine System Integrated Deep Learning Model

The Human-Machine Identity Verification System based on Deep Learning offers a robust and automated approach to identity verification, leveraging the power of deep learning algorithms to enhance accuracy and security. This paper focused on the biometric-based authentical scheme with Biometric Recognition for the Huma-Machinary Identification System. The proposed model is stated as the Two-Factor Biometric Authentication Deep Learning (TBAuthDL). The proposed TBAuthDL model uses the iris and fingerprint biometric data for authentication. TBAuthDL uses the Weighted Hashing Cryptographic (WHC) model for the data security. The TBAuthDL model computes the hashing factors and biometric details of the person with WHC and updates to the TBAuthDL. Upon the verification of the details of the assessment is verified in the Human-Machinary identity. The simulation analysis of TBAuthDL model achieves a higher accuracy of 99% with a minimal error rate of 1% which is significantly higher than the existing techniques. The performance also minimizes the computation and processing time with reduced complexity.

A novel negative selection algorithm with hyper-rectangle detectors based on full coverage of state space for anomaly detection

The traditional negative selection algorithm with a randomly generated hypersphere detector is unable to satisfy the development needs of continuous learning due to the inherent defects of the detector. This paper proposes a novel negative selection algorithm for hyper-rectangle detectors that overcomes the shortcomings of randomly generated hyper-sphere detectors and lays the foundation for a negative selection algorithm with continuous learning capability. It uses self-sample clusters of equal-sized hypercubes instead of self-samples for training. The hyper-rectangle detectors are generated by cutting the nonself-space along the boundary of the self-sample clusters. The state space is covered without overlapping each other by self-sample clusters and detectors. The anomaly detection performance of the proposed method was demonstrated using Iris data, vowel recognition data (Vowel), and Wisconsin Breast Cancer (BCW) data. The experimental results show that the proposed method outperforms other artificial immune algorithms and clustering algorithms under the same parameter conditions.

ASSESSMENT OF THE IRIS RECOGNITION USING EDGE DETECTION ALONG BENCHMARK DATASETS

An individual's identity has grown more essential for individuals to satisfy modern corporate the community's higher safety standards. Iris represents one of the finest and most precise biometric systems now in utilization, with multiple edge detection methods used. As a consequence, understanding the different kinds of edge detection algorithmic methods which are now under operation is crucial. Three edge detection methods are used to be assessed on iris data during the present research. These methodologies are carried out using the MATLAB environment, which provides the evaluation with verified results. CASIA and MMU are the datasets that are used for this purpose. In contrast, the results reveal that the canny edge detection method works quite efficiently. Visual quality is crucial in vision-based recognition of objects. The present research assesses the visual aesthetic using various quality indicators such as PSNR and MSE. When compared to PSNR and MSE, however, MSE is recognized as the best image quality statistic.

Efficient private key generation from iris data for privacy and security applications

Private key generation based on biometric data is gaining popularity in several cryptographic applications. Some key generation approaches using fingerprint and face data are known for this purpose. The existing approaches, in general, follow direct features from raw data, which may have an issue of revealing users’ biometric data. Further, iris biometrics being a better source for stable and secure key generation, is yet to be explored. In this work, an iris biometric-based key generation approach is proposed. In the proposed method, first, an ensemble L0 Gradient Minimization and an edge-preserving filter are used to extract iris structure from the iris images. Secondly, the iris texture data is normalized, and the consistent region is selected based on a novel statistical method. Thirdly, feature vectors are generated using the ensemble local space-filling curve descriptors and their variants. Fourth, a discriminant feature vector is generated using hybrid feature selection and neighbourhood component analysis. Finally, an interval-based encoding scheme is used to generate a key from the discriminant feature vector. To substantiate the efficacy of the proposed approach, extensive experiments have been carried out with near-infrared images, visible wavelength images, at-a-distance, and on-move images. Further, distinctiveness, dissimilarity, randomness, and security analyses have been carried out to validate the resilience of keys against different attacks. More significantly, a unique key can be generated dynamically, that is, from an online image. The proposed approach is applicable to different cryptographic applications, such as data storage security, remote authentication protocol, blockchain system security, etc.

Learning by non-interfering feedback chemical signaling in physical networks

Both non-neural and neural biological systems can learn. So rather than focusing on purely brain-like learning, efforts are underway to study learning in physical systems. Such efforts include equilibrium propagation (EP) and coupled learning (CL), which require storage of two different states---the free state and the perturbed state---during the learning process to retain information about gradients. Here, we propose a learning algorithm rooted in chemical signaling that does not require storage of two different states. Rather, the output error information is encoded in a chemical signal that diffuses into the network in a similar way as the activation/feedforward signal. The steady-state feedback chemical concentration, along with the activation signal, stores the required gradient information locally. We apply our algorithm using a physical, linear flow network and test it using the Iris data set with 93% accuracy. We also prove that our algorithm performs gradient descent. Finally, in addition to comparing our algorithm directly with EP and CL, we address the biological plausibility of the algorithm.

Properties of dust in the North - East part of Perseus Cloud within the open cluster IC 348 using data from IRIS and AKARI

In this work, we have studied the dust properties of the North-East part of Perseus cloud having a size of 0.5o × 0.5o located at RA (ICRS): 56.14o DEC (ICRS): +32.15o, within the Open Cluster IC 348, using IRIS and AKARI data. An isolated region of size 0.39o×0.23o in IRIS and 0.34o×0.16o in AKARI data is detected within it. The infrared fluxes extracted using Aladin v11.0 are used to study the dust temperature and dust mass. The distance of the selected dust structure is calculated using Gaia EDR3, which is 309.98 pc. The infrared flux density is found to be increased for long-wavelength data. The average dust color temperature calculated from short-wavelength IRIS data is more than that calculated from long-wavelength AKARI data, which are 26.34 K±0.11 K for IRIS and 17.63 K±0.02 K for AKARI data. The mass of dust within the entire dust structure is 0.06 Mʘ for IRIS data and 37.44 Mʘ for AKARI data. Jeans mass for isolated region gives contradictory results in two surveys, for IRIS survey total mass is smaller than Jeans mass but for AKARI survey the total mass is larger than the Jeans mass. A good correlation between infrared fluxes is noticed for linear regression. The study background sources observed from the SIMBAD database explore the large number of stars, X-ray sources, YSOs, etc., embedded within the dust structure; some of them are responsible for dust heating and some for the contribution of dust mass. The contour map shows the identical distribution between infrared fluxes, dust color temperature and Planck’s function and dissimilar distribution between dust mass and visual extinction in IRIS and AKARI data. The high temperature at the central region suggests that the core of dust structure is thermally active, radiating the large thermal radiation in comparison to the outer region.

Comparison of aftershock behavior of the flores sea 12 december 1992 and 14 december 2021

<span lang="EN-US">Over 30 years, the Flores region experienced earthquakes with Mw >7 twice, namely in 1992 and 2021 at shallow depths <60 km. However, the behavior of the aftershocks accompanying the quake has yet to be studied in detail. This research aims to compare the behavior of aftershocks after the main earthquake. This research uses three-lifetime analysis functions, namely, Wilber3, Mustang Web Browser, and comparison of Probability Density Function Hazard. The data used comes from IRIS data and BMKG Online Data. Through graphical analysis, it is known that the data for the time difference for aftershocks follow the RMS distribution, indicating that the Flores Sea aftershock on 12 December 1992 occurred with high intensity on the 3<sup>rd</sup> and the 6<sup>th</sup> day after the main earthquake. Meanwhile, the aftershocks in the Flores Sea on 14 December 2021 tended to occur with a more even intensity during the observation period, but the power returned to high on the 7<sup>th</sup> day</span>

Double-Center-Based Iris Localization and Segmentation in Cooperative Environment with Visible Illumination.

Iris recognition has been considered as one of the most accurate and reliable biometric technologies, and it is widely used in security applications. Iris segmentation and iris localization, as important preprocessing tasks for iris biometrics, jointly determine the valid iris part of the input eye image; however, iris images that have been captured in user non-cooperative and visible illumination environments often suffer from adverse noise (e.g., light reflection, blurring, and glasses occlusion), which challenges many existing segmentation-based parameter-fitting localization methods. To address this problem, we propose a novel double-center-based end-to-end iris localization and segmentation network. Different from many previous iris localization methods, which use massive post-process methods (e.g., integro-differential operator-based or circular Hough transforms-based) on iris or contour mask to fit the inner and outer circles, our method directly predicts the inner and outer circles of the iris on the feature map. In our method, an anchor-free center-based double-circle iris-localization network and an iris mask segmentation module are designed to directly detect the circle boundary of the pupil and iris, and segment the iris region in an end-to-end framework. To facilitate efficient training, we propose a concentric sampling strategy according to the center distribution of the inner and outer iris circles. Extensive experiments on the four challenging iris data sets show that our method achieves excellent iris-localization performance; in particular, it achieves 84.02% box IoU and 89.15% mask IoU on NICE-II. On the three sub-datasets of MICHE, our method achieves 74.06% average box IoU, surpassing the existing methods by 4.64%.

Research on Data Classification Method of Optimized Support Vector Machine Based on Gray Wolf Algorithm

The data classification method based on support vector machine (SVM) has been widely used in various studies as a non-linear, high precision, and good generalization ability machine learning method. Among them, the kernel function and its parameters have a great impact on the classification accuracy. In order to find the optimal parameters to improve the classification accuracy of SVM, this paper proposes a data multi-classification method based on gray wolf algorithm optimized SVM(GWO-SVM). In this paper, the iris data set is used to test the performance of GWO-SVM, and the classification result is compared with those based on genetic algorithm (GA), particle swarm optimization (PSO) and the original SVM model. The test results show that the GWO-SVM model has a higher recognition and classification accuracy than the other three models, and has the shortest running time, which has obvious advantages and can effectively improve the classification accuracy of SVM. This method has practical significance in image classification, text classification, and fault detection.

Variational quantum semi-supervised classifier based on label propagation

Label propagation is an essential semi-supervised learning method based on graphs, which has a broad spectrum of applications in pattern recognition and data mining. This paper proposes a quantum semi-supervised classifier based on label propagation. Considering the difficulty of graph construction, we develop a variational quantum label propagation (VQLP) method. In this method, a locally parameterized quantum circuit is created to reduce the parameters required in the optimization. Furthermore, we design a quantum semi-supervised binary classifier based on hybrid Bell and Z bases measurement, which has a shallower circuit depth and is more suitable for implementation on near-term quantum devices. We demonstrate the performance of the quantum semi-supervised classifier on the Iris data set, and the simulation results show that the quantum semi-supervised classifier has higher classification accuracy than the swap test classifier. This work opens a new path to quantum machine learning based on graphs.

A medoid-based deviation ratio index to determine the number of clusters in a dataset

Most existing methods of determining the number of groups apply to particular data types or are calculated based on the distance matrix for all object pairs. In this paper, we propose a medoid-based Deviation Ratio Index (DRI) to determine the number of clusters. The DRI is calculated based on the distance matrix for each object to k final medoids. These final medoids are produced by the block-based k-medoids algorithm (BlockD-KM). We choose a specific transformation and a suitable distance for certain variables before executing the BlockD-KM. We illustrated the detailed stages of DRI on secondary data in the 2022 environmental index of Asia Pacific countries, so that they are easy to reproduce. We use eight real datasets, namely Breast Cancer, Heart Disease, Iris, Wine, Soybean, Ionosphere, Vote, and Credit Approval data, to validate the DRI method. We compare the DRI method with the Calinski-Harabaz (CH) and the Silhouette index. The experimental results show that the DRI is 100% correct in predicting the number of clusters. While the CH index correctly predicts 62.5% and the Silhouette index of 75%. We also generated three kinds of artificial data to evaluate the proposed method, and 76.7% of the experiments were predicted correctly.•The medoid-based deviation ratio index aids the researcher in determining the number of clusters•The DRI method applicable to any medoids-based partitioning algorithm•This method is suitable for all data types (categorical, numerical, and mixed)

자가 진단을 위한 스마트폰 적용 LED 홍채 렌즈 연구

Digital health services tend to go beyond simple health management and auxiliary means to advanced disease monitoring. The use of apps related to particular diseases such as diabetes and the circulatory system is a large increasing every year. In particular, there is connected to the brain and nervous system tissues so it is possible to diagnose the disease state of the related tissue from all the changes in the iris image. Although an iris glasses camera module was developed in the previous study, there was a limit in user convenience in the method of acquiring images using the USB communication method. In addition, if the disease is read with an opaque iris image when taking an iris photograph with a smartphone, the patient’s disease may be missed as a result of no disease, so it is important to obtain a clear iris image. This research proposed an iris recognition lens implementation of a small and can be easily carried and designed considering the safety of patients with high-definition iris image capture. In this study, an LED iris lens was designed and manufactured can iris recognition to show the crown of the autonomic nervous system. It is necessary to verify whether the developed LED iris lens can perform to show the crown of the autonomic nervous system. Measurements are classfied in two ways iris clarity through smartphones, iris clarity through Test Tool, and the number of attempts to secure a clear iris image. The development lange of the Test Tool is Python, and the Average Square Shrinking algorithm was applied to detect autonomic neuropathy in the iris. When verifying the iris, the iris data set provided by CASIA(China’s Biological Research Institute)was used as reference data. As a result of the test, The conventional app for iris recognition can obtain at most 12% of clarity which is not enough to diagnosis the diseases while using the LED iris lens on the smartphone gives 91% of the iris brightness. Additionally, the number of trials for iris recognition was only once while the conventional method needs at least 8 times to get a clear iris image. As a result of verifying the iris image by developing an LED iris lens through this study, it was possible to secure a clear iris image at the level of disease diagnosis. A clear iris reduces diagnostic errors and contributes sufficiently to patient safety. There is a lot of interest in mobile health apps for self-diagnosis. However, it is difficult to find useful App. for self-diagnosis. After applying the LED iris lens developed through this study, it is judged that it will be of practical help to disease prevention if the health status of the iris is checked periodically.

Quaternion model of Pythagorean fuzzy sets and its distance measure

Pythagorean fuzzy set (PFS), one of several non-standard fuzzy sets, introduces a number pair a,b that satisfies the condition a2+b2≤1 to express membership grade. The above expression suggests that PFS provides a more extensive description space for describing fuzzy information, thereby attracting much attention in scientific research and engineering practice. Many studies have attempted to further exploit the potential of PFS in a reasonable manner. Consequently, this paper proposes a quaternion model of Pythagorean fuzzy set (QPFS). In QPFS, membership, non-membership, and hesitation function are expressed in the form of quaternion. A major advantage of QPFS over PFS is that its description space of fuzzy information stretches from the real plane to the hypercomplex plane. This is useful for capturing the composite features of fuzzy information as well as describing multi-dimensional fuzzy information. We then define the basic logical operations of QPFS, including complement, union, and intersection, and also derive the properties of these operations. As an additional contribution, this paper presents a distance measure (QPFSD) in the quaternion model of Pythagorean fuzzy sets (QPFSs). QPFSD is a strict distance measure that satisfies the axioms of distance measure, i.e., nonnegativity, symmetry, and triangle inequality. When QPFSs collapse into Pythagorean fuzzy sets (PFSs), QPFSD becomes the distance measure of real space. It has been demonstrated through some numerical examples that QPFSD is an effective method for measuring difference between QPFSs. Furthermore, we apply QPFSD to domains such as expert evaluation and data-driven environments in the QPFS framework. Experimental results related to expert evaluation suggest that QPFSD and the models involving it can provide an intuitive decision-making. Experimental results based on Iris data set demonstrate that with the increase of the number of training sets, the recognition rate of methods associated with QPFSD also increases for unknown targets. With the help of iris data, based on QPFSD, we investigate the performance of QPFS, CPFS, PFS and FS, respectively. The average recognition rate associated with QPFS is the highest among these. Through the use of iris dataset experiments and numerical example, this paper compares QPFSD and other distance measures in the CPFS and PFS frameworks, respectively. As a result of the dataset experiment, the recognition of methods pertaining to QPFSD is highest within CPFS and PFS frameworks. The numerical experiments indicate that QPFSD is highly sensitive to the differences between fuzzy information compared with other distance measures.