A Robust Authentication System Using Multiple Biometrics

Sentiment analysis extracts the emotions expressed in text and has been employed in many fields including politics, elections, movies, retail businesses and in recent years microblogs to understand, track and control the human sentiments or reactions toward products events or ideas. Nevertheless challenges such as different styles of writing, use of negation and sarcasm, existence of spelling mistakes, invention of new words etc. provide obstacle in the correct classification of sentiments. This paper provides an ensemble of classifiers framework for sentiment analysis. The proposed weighted majority voting ensemble method combines six models including Naïve Bayes, Logistic Regression, Stochastic Gradient Descent, Random Forest, Decision Tree and Support Vector Machine to form a single classifier. Weights of the individual classifiers of the ensemble are chosen as accuracy or Fl-score by optimizing their performance. This approach combines models based on the simple majority voting as opposed to the one based on weighted majority voting. Additionally, a comparison is drawn among these six individual classifiers to evaluate their performance. The proposed ensemble model is tested on some existing sentiment datasets, including SemEval 2017 Task 4A, 4B and 4C. The results demonstrate that the Logistic Regression classifier is optimal as compared to other individual classifiers. Furthermore, the proposed ensemble weighted majority voting classifier with the six individual classifiers performs better compared to the simple majority voting and all independent classifiers.

Relevant with some important subjects like target recognition, sensor fusion systems can be considered as one of the main issues highlighting here. Environmental condition, target characteristic and sensor efficiency are three parameters which can impress on sensor value in target recognition so for recognizing targets, a group of sensors which have more recognition rates, must be selected intelligently. Utilizing many sensors to acquire the highest object recognition rate would have extra cost and decrease energy of mobile sensors rapidly. Therefore make a tradeoff between sensor numbers and object recognition rate would be imperatively. This paper attempts to design a multi objective optimization service by using optimization algorithm and neural network. This service specifies highest recognition rate for each distinct sensor numbers. We propose multi objective optimization algorithm to help accessing the best sensory configuration for a definite environment regarding to the environmental conditions, sensors performance, and object features. Our multi objective optimization algorithm has two functions. Genetic algorithm is used to perform as one of functions to specify object recognition rates of each sensor group. Neural network is used to perform as fitness function of each genetic algorithm chromosomes. Another function is sensor numbers determinant. Highest recognition rate and lowest sensor numbers are two objects which multi objective optimization algorithm wants to make a balance between them. We define 500 different scenarios for 6 different sensors in different conditions. Object recognition rate of each sensor is collected. These rates are used for neural networks training process. By defining new scenario and run multi objective optimization algorithm in this scenario, this algorithm makes a Pareto front between sensor numbers and object recognition rate. Finally this algorithm finds, by distinct numbers of sensor, which sensors by which recognition ability must be used to reach the highest recognition rate.

Biometric systems use personal biological or behavioural traits that can uniquely characterise an individual but this uniqueness property also becomes its potential weakness when the template characterising a biometric trait is stolen or compromised. To this end, we consider two strategies to improving biometric template protection and performance, namely, (1) using multiple privacy schemes and (2) using multiple matching algorithms. While multiple privacy schemes can improve the security of a biometric system by protecting its template; using multiple matching algorithms or similarly, multiple biometric traits along with their respective matching algorithms, can improve the system performance due to reduced intra-class variability. The above two strategies lead to a novel, ensemble system that is derived from multiple privacy schemes. Our findings suggest that, under the worst-case scenario evaluation where the key or keys protecting the template are stolen, multi-privacy protection scheme can outperform a single protection scheme as well as the baseline biometric system without template protection.

An investigation of ensemble combination schemes for genetic programming based hyper-heuristic approaches to dynamic job shop scheduling

Plant diseases pose significant economic challenges and may lead to ecological consequences. Although plant pathologists have a significant ability to diagnose plant diseases, rapid, accurate, and early diagnosis of plant diseases by intelligent systems could improve disease control and management. This study evaluates six efficient classification models (classifiers) based on deep learning to detect common tomato diseases by analyzing symptomatic patterns on leaves. Additionally, group learning techniques, including simple and weighted majority voting methods, were employed to enhance classification performance further. Six tomato leaf diseases, including Pseudomonas syringae pv. syringae bacterial spot, Phytophthora infestance late blight, Cladosporium fulvum leaf mold, Septoria lycopersici Septoria leaf spot, Corynespora cassiicola target spot, and Alternaria solani early blight, as well as healthy leaves, resulting in a total of seven classes, were utilized for the classification. Deep learning models, such as convolutional neural networks (CNNs), GoogleNet, ResNet-50, AlexNet, Inception v3, and MobileNet, were utilized, achieving classification accuracies of 65.8%, 84.9%, 93.4%, 89.4%, 93.4%, and 96%, respectively. Furthermore, applying the group learning approaches significantly improved the results, with simple majority voting achieving a classification accuracy of 99.5% and weighted majority voting achieving 100%. These findings highlight the effectiveness of the proposed deep ensemble learning models in accurately identifying and classifying tomato diseases, featuring their potential for practical applications in tomato disease diagnosis and management.

Boosting is recognized as one of the most successful techniques for generating classifier ensembles. Typically, the classifier outputs are combined by the weighted majority vote. The purpose of this study is to demonstrate the advantages of some fuzzy combination methods for ensembles of classifiers designed by Boosting. We ran two-fold cross-validation experiments on six benchmark data sets to compare the fuzzy and nonfuzzy combination methods. On the "fuzzy side" we used the fuzzy integral and the decision templates with different similarity measures. On the "nonfuzzy side" we tried the weighted majority vote as well as simple combiners such as the majority vote, minimum, maximum, average, product, and the Naive-Bayes combination. In our experiments, the fuzzy combination methods performed consistently better than the nonfuzzy methods. The weighted majority vote showed a stable performance, though slightly inferior to the performance of the fuzzy combiners.

Biometric identification is thought to be less vulnerable to fraud and forgery than are traditional forms of identification. However biometric identification is not without vulnerabilities. In a 'spoofing attack' an artificial replica of an individual's biometric trait is used to induce a system to falsely infer that individual's presence. Techniques such as liveness-detection and multi-modality, as well as the development of new and emerging modalities, are intended to secure biometric identification systems against such threats. Unlike biometrics in general, the societal and ethical issues raised by spoofing and anti-spoofing techniques have not received much attention. This paper examines these issues.

Recent works have shown that multimodal biometric systems can be evaded by spoofing only a single biometric trait. In this paper, we propose a method to evaluate the robustness of such systems against spoofing attacks, when score-level fusion rules are used. The aim is to rank several score-level fusion rules, to allow the designer to choose the most robust one according to the model predictions. Our method does not require to fabricate fake biometric traits, and allows one to simulate different possible spoofing attacks using the information of genuine and impostor distributions. Reported results, using data set containing realistic spoofing attacks, show that our method can rank correctly score-level fusion rules under spoofing attacks.

Annotating carious lesions on images is challenging. For artificial intelligence (AI) applications, the aggregation of heterogeneous multi-examiner annotations into one single annotation (e.g. via majority voting, MV) is usually needed. We assessed different aggregation strategies for multi-examiner annotations of primary proximal carious lesions on orthoradial radiographs and Near-Infrared Light Transillumination (NILT) images. A total of 1007 proximal surfaces from 522 extracted posterior teeth were assessed by five dentists. Histological analysis provided the gold standard. Surfaces were classified as (1) sound, (2) enamel lesion or (3) dentin lesion. Four label aggregation strategies - MV, Weighted Majority Voting (WMV), Dawid-Skene (DS), and multi-annotator competence estimation (MACE) - were applied to unimodal (radiographs, NILT) and multimodal (combined) datasets. The area under the receiver operating characteristic curve (AUROC) was the primary outcome metric. According to the gold standard, 637 (63 %) surfaces were sound, 280 (28 %) showed carious lesions limited to the enamel, and 90 (9 %) showed lesions extending into the dentin. For radiographs, aggregation using MACE outperformed MV, WMV and DS significantly across all lesion depths (p < 0.002). For NILT, MACE significantly outperformed MV across all lesion depths (p < 0.001) and DS for enamel and dentin lesions (p ≤ 0.002). In the multimodal dataset, DS outperformed the other label aggregation strategies across all lesion depths significantly (p < 0.05). The commonly applied MV may be suboptimal. There is a need for informed application of specific aggregation strategies, depending on the dataset characteristics. Most AI applications for dental image analysis are trained on a single annotation, usually resulting from aggregated multi-examiner annotations of each image. However, since these annotations are usually aggregated in an in vivo setting where no definitive ground truth is available, the choice of aggregation strategy plays a crucial role.

The Biometrics denotes the metrics that are connected to the features of the human being. Biometric traits are normally used to report the individuals. Biometric determiners are normally divided as behavioural or the physiological identifiers, specifically related to the behaviour of person(s) or the shape, respectively, and using the determiners such as the face, iris or retina, finger–knuckle prints, etc., help to decrease the chances of any system to be compromised. Identity verification of individual(s) is treated as the basic principle in all the organizations, be that a governmental organization, semi-government organization, legal or forensic operations, or in the civilian applications. This principle of this research is to make the use of the various Biometrics traits together, which helps in overcoming the problems that are usually raised in traditional-recognition systems, thus making the recognition systems well-protected from the spoofing attacks. The suggested work aims to make the attacks on the recognition system challenging task, and adding the preventive and detective measures against the spoof attacks. The main purpose of this work is to make sure that the fundamental principle of security, that is integrity, is not compromised.

Finger-vein recognition is a biometric technology, which makes use of the pattern of veins in the skin of fingers. Finger-vein recognition is widely employed owing to its high recognition rate and ease of use. A spoof detection method is essential for identifying spoof attacks in biometric-recognition systems. Studies have primarily investigated methods for detecting spoof attacks that utilize image printing or replay using fake finger-vein images. However, no studies have investigated the detection of spoof attacks based on fake finger-vein images generated through a generative adversarial network (GAN), which is commonly implemented in deep-learning technologies. To solve this problem, we newly propose a spoof attack detection method robust against GAN-based fake finger-vein image generation. We verified the use of spoof finger-vein images generated through cycle-consistent adversarial networks (CycleGAN) to attack conventional finger-vein recognition systems and used this finding as the motivation for our research. We also propose a finger-vein spoof-detection method that employs score fusion of ensemble networks based on support vector machine using adaptive moment estimation (Adam) with a sharpness-aware minimization (SAM) optimizer. We conducted experiments using two open databases, ISPR and Idiap, to evaluate the proposed framework. The average classification error rates (ACER) of spoof detection using the proposed method with ISPR and Idiap were 0.37% and 0.23%, respectively. This result confirms the better performance by our method than conventional state-of-the-art methods.

Toxic plants have been a major threat to public health in China. However, identification and tracing of poisoned species with traditional methods are unreliable due to the destruction of plant morphology by cooking and chewing. DNA barcoding is independent of environmental factors and morphological limitations, making it a powerful tool to accurately identify species. In our study, a total of 83 materials from 26 genera and 31 species of 13 families were collected and 13 plant materials were subjected to simulated gastric fluid digestion. Four markers (rbcL, trnH-psbA, matK, and ITS) were amplified and sequenced for all untreated and mock-digested samples. The effectiveness of DNA barcoding for the identification of toxic plants was assessed using Basic Local Alignment Search Tool (BLAST) method, PWG-Distance method, and Tree-Building (NJ) method. Except for the matK region, the amplification success rate of the remaining three regions was high, but the sequencing of trnH-psbA and ITS was less satisfactory. Meanwhile, matK was prone to be more difficult to amplify and sequence because of simulated gastric fluid. Among the three methods applied, BLAST method showed lower recognition rates, while PWG-Distance and Tree-Building methods showed little difference in recognition rates. Overall, ITS had the highest recognition rate among individual loci. Among the combined loci, rbcL + ITS had the highest species recognition rate. However, the ITS region may not be suitable for DNA analysis of gastric contents and the combination of loci does not significantly improve species resolution. In addition, identification of species to the genus level is sufficient to aid in the clinical management of most poisoning events. Considering primer versatility, DNA sequence quality, species identification ability, experimental cost and speed of analysis, we recommend rbcL as the best single marker for clinical identification and also suggest the BLAST method for analysis. Our current results suggest that DNA barcoding can rapidly identify and trace toxic species and has great potential for clinical applications. In addition, we suggest the creation of a proprietary database containing morphological, toxicological and molecular information to better apply DNA barcoding technology in clinical diagnostics.

To improve weak classifiers bagging and boosting could be used. These techniques are based on combining classifiers. Usually, a simple majority vote or a weighted majority vote are used as combining rules in bagging and boosting. However, other combining rules such as mean, product and average are possible. In this paper, we study bagging and boosting in Linear Discriminant Analysis (LDA) and the role of combining rules in bagging and boosting. Simulation studies, carried out for two artificial data sets and one real data set, show that bagging and boosting might be useful in LDA: bagging for critical training sample sizes and boosting for large training sample sizes. In contrast to a common opinion, we demonstrate that the usefulness of boosting does not directly depend on the instability of a classifier. It is also shown that the choice of the combining rule may affect the performance of bagging and boosting.KeywordsLinear Discriminant AnalysisMajority VoteGeneralization ErrorTraining ObjectBearing FaultThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

A new privilege of biometrics help to reduce the stress of user, Which comes along with the traditional access methods of passwords and token. Using the biometrics limitations and weaknesses can be knocked out. However, biometrics has raise privacy risks and new security since they cannot be easily revoked. Due to the spoofing attack on biometrics. Thus, to protect biometric traits against spoofing attack a multimodal biometric jammer scheme for the security enhancement have been developed and suggested in this paper. Firstly, we analyze why the multimodal biometric system have attracted attention for high security-demanding schemes. Secondly, security of biometric system is increasing and prevented it from spoofing attack developing a machine learning system model. We show that these machine learning algorithms perform pre-processing of biometric traits images. Further we analyze user identification with the increase precision and reliability using biometric features. Where feature extraction of each one trait of biometric is done and then all features are concatenation to get a single feature. With the aid of machine learning classifier using extracted features the algorithm predict the result of the system.

Digital content refinement by accumulation of defectively reliable publications and majority voting over a network