Finger knuckle print recognition for personal authentication based on relaxed local ternary pattern in an effective learning framework

Biometric methods are among the safest and most secure solutions for identity recognition and verification. One of the biometric features with sufficient uniqueness for identity recognition is the finger knuckle print (FKP). This paper presents a new method of identity recognition and verification based on FKP features, where feature extraction is combined with an entropy-based pattern histogram and a set of statistical texture features. The genetic algorithm (GA) is then used to find the superior features among those extracted. After extracting superior features, a support vector machine-based feedback scheme is used to improve the performance of the biometric system. Two datasets called Poly-U FKP and FKP are used for performance evaluation. The proposed method managed to achieve 94.91{%} and 98.5{%} recognition rates on the Poly-U FKP and FKP datasets and outperformed all of the existing methods in this respect. These results demonstrate the potential of this method as a simple yet effective solution for FKP-based identity recognition.

Online finger-knuckle-print verification for personal authentication

Multimodal biometric systems represent a significant advancement in biometric authentication technology by integrating multiple modalities to enhance accuracy and security. Our proposed system combines electrocardiogram (ECG), fingerprint, and finger knuckle print (FKP) modalities to achieve improved authentication performance, especially suited for high-security applications. The system first uses ECG for liveness detection, ensuring that only genuine users proceed, followed by fingerprint and FKP for authentication. Three individual Siamese Neural Networks were developed, each optimized to extract distinct features from each modality. Data quality was enhanced using preprocessing methods such as noise reduction and normalization, along with data augmentation strategies to improve model robustness. The system was evaluated using a balanced dataset of 50 samples per biometric modality. Individual average accuracies reached 99.54% for ECG, and 100% for both fingerprint and FKP considering 90 subjects. Using weighted average score-level fusion with a priority on fingerprint and FKP, the system achieved an overall accuracy of 99.80%, with a False Acceptance Rate (FAR) of 0.20%, False Rejection Rate (FRR) of 0.21%, Equal Error Rate (EER) of 0.20%, and an F1-Score of 99.80%. These results demonstrate the system's resilience to spoofing and robustness against data variability, offering a highly secure authentication solution with practical applicability in cloud-based high-security environments, such as banking and healthcare.

Finger Knuckle Print as Unimodal Fuzzy Vault Implementation

This paper presents anovel method for texture classification and biometric authentication based on a descriptor called the weighted mean-based patterns (WEM). The proposed descriptor has been developed for extracting texture features from a large dataset of hand images, which has been created by the authors. The method uses the distinctive features of finger knuckle print (FKP) for hand image retrieval, which can be used in biometric identity recognition systems. The proposed method also includes a feature selection step for eliminating less important patterns and a weighted distance measure for quantifying the similarity of images. The method uses the support vector machine (SVM) for the classification stage. The proposed method has been tested on the FKP image dataset to evaluate the image retrieval performance, and also on Brodatz, Vistex, and Stex datasets to evaluate the performance for texture classification. Higher performance of the proposed method is demonstrated through comparison with other methods. The proposed method is shown to be sufficiently precise for a variety of applications, including identity recognition and classification.

1st International Workshop on Emerging Techniques and Challenges for Hand-Based Biometrics, ETCHB 2010, Istanbul, 22-22 August 2010

A framework is proposed for how bilingual learners develop knowledge, language, and mathematics literacy. The framework centers on principles of learning, effective pedagogy, and second language acquisition theories, and these elements are incorporated in a mathematics lesson depicted in this article.

In this paper, multi-resolution analysis of two edge-texture based descriptors, Discriminative Robust Local Binary Pattern (DRlbp) and Discriminative Robust Local Ternary Pattern (DRltp), are proposed for the determination of mammographic masses as benign or malignant. As an extension of Local Binary Pattern (LBP) and Local Ternary Pattern (LTP), DRlbp and LTP-based features overcome the drawbacks of these features preserving the edge information along with texture. With the hypothesis that multi-resolution analysis of these features for different regions related to mammaographic masses with wavelet transform will capture more discriminating patterns and thus can help in characterizing masses. In order to evaluate the efficiency of the proposed approach, several experiments are carried out using the mini-MIAS database where a 5-fold cross validation technique is incorporated with Support Vector Machine (SVM) on the optimal set of features obtained via stepwise logistic regression method. An area under the receiver operating characteristic (ROC) curve ([Formula: see text] value) of 0.96 is achieved with DRlbp attributes as the best performance. The superiority of the proposed scheme is established by comparing the obtained results with recently developed other competing schemes.

Various biometric traits are available but recently finger knuckle image has attracted great attention to biometric research community due to its potentiality and ease of use. The performance of any biometric system heavily depends on the accuracy of feature detection and robustness of feature description. A number of feature descriptors are available but selection is being determined by the type of application such as image retrieval, biometric system, remote sensing etc. This paper proposed a biometric system using leading descriptor for finger knuckle biological trait image recognition and also compare the proposed system with existing leading state-of-art finger knuckle print recognition. The recognition performance is measured by some standard evaluation protocol such as Equal Error Rate (EER), Decidability index, Computation cost, Zero-score imposter probability, Zero-score genuine probability, Receiver Operating Characteristic (ROC), Detection Error Trade-off (DET) over PolyU Finger Knuckle benchmark database. The experimental results show that, the performance of SURF, KAZE and ORB are comparable and are better as compared to BRISK and MSER descriptor. The ERR of 0.0010% is obtained with ORB descriptor while the Decidability index of 6.4645 is obtained for KAZE. The minimum Computational cost of 0.1442 s is obtained for SURF as compares to other of its class.

In this paper, we present a novel image representation approach for classifying emphysema in computed tomography (CT) images of the lung. Our proposed method extends rotation invariant uniform local binary pattern (RIULBP) and local ternary pattern (LTP), which are extensively used in a variety of computer vision applications, into rotation invariant uniform local ternary pattern (RIULTP) with a human perception principle: Weber's law. In addition, by integrating the upper pattern and the lower pattern of the Weber-based RIULTP (WRIULTP), we further put forward the joint Weber-based rotation invariant uniform local ternary pattern (JWRIULTP), which allows for a much richer representation and also takes the comprehensive information of the image into account. The proposed methods are tested on the Outex database (texture database) and the Bruijne and S⊘rensen database (emphysema database). The results show the superiority of the proposed approaches to the state-of-the-art techniques for emphysema classification including rotation invariant local binary pattern (RILBP) and texton-based approach.

This chapter outlines key findings of cognitive and developmental psychology which could be used as a theoretical framework to guide the design and research of e-learning applications. The chapter consists of two main sections. The first section presents the basic cognitive mechanisms and their development, while the second part discusses how our knowledge of the cognitive system can guide the design of computer-based learning environments. It is proposed that the human mind is organized in three levels: two of them are general-purpose mechanisms and processes and one consists of domainspecific structures of knowledge representation and problem solving processes. These three levels are associated with an effective learning framework. The suggested framework describes the basic points that any designer of e-learning environments must consider. The second section of the chapter discusses e-learning, connecting it to knowledge about the human cognitive system. In this section, we first present the conceptual bounds of the e-learning construct and discuss how cognitive theories of learning should guide the design of technologically-enhanced learning environments. It is proposed that a synergy between the study of human cognition and the design of e-learning applications is required for effectively understanding both fields.

Scaling object taxonomies is one of the important steps toward a robust real-world deployment of recognition systems. We have faced remarkable progress in images since the introduction of the LVIS benchmark. To continue this success in videos, a new video benchmark, TAO, was recently presented. Given the recent encouraging results from both detection and tracking communities, we are interested in marrying those two advances and building a strong large vocabulary video tracker. However, supervisions in LVIS and TAO are inherently sparse or even missing, posing two new challenges for training the large vocabulary trackers. First, no tracking supervisions are in LVIS, which leads to inconsistent learning of detection (with LVIS and TAO) and tracking (only with TAO). Second, the detection supervisions in TAO are partial, which results in catastrophic forgetting of absent LVIS categories during video fine-tuning. To resolve these challenges, we present a simple but effective learning framework that takes full advantage of all available training data to learn detection and tracking while not losing any LVIS categories to recognize. With this new learning scheme, we show that consistent improvements of various large vocabulary trackers are capable, setting strong baseline results on the challenging TAO benchmarks. KeywordsLarge vocabularyVideo object detection and tracking

This paper investigates the potential of gait-based authentication for securing virtual environments, specifically within the Metaverse. With the growing need for reliable and secure identity verification in virtual spaces, traditional authentication methods, such as passwords or PINs, have proven insufficient. In contrast, biometric authentication systems, including gait analysis, provide a more secure and user-friendly alternative by leveraging unique physiological and behavioral traits for identity verification. This research applies machine learning algorithms—Random Forest and Support Vector Machine (SVM)—to gait data for distinguishing between authentic users and imposters. The dataset consists of 1,000 simulated gait samples with 16 features, such as stride length, step frequency, joint angles, and ground reaction forces (GRF). After performing exploratory data analysis (EDA), including feature distribution visualization and correlation analysis, two models were trained on the data. The Random Forest model outperformed the SVM model, achieving an accuracy of 56% and a recall of 76%, indicating its effectiveness in identifying authentic users. Despite the promising results, both models showed only marginal improvement over random guessing, highlighting the need for further optimization. This study contributes to the growing body of research on gait-based biometric systems by demonstrating their potential as a viable method for identity verification in virtual environments. It also identifies the most important gait features, such as step frequency, cadence variability, and knee joint angle, that significantly contribute to the classification process. Future research should explore advanced deep learning techniques and the integration of multimodal biometric systems to enhance the performance and reliability of gait-based authentication.

Ceria (CeO(2)) and ceria-based composite materials, especially Ce(1-x)Zr(x)O(2) solid solutions, possess a wide range of applications in many important catalytic processes, such as three-way catalysts, owing to their excellent oxygen storage capacity (OSC) through the oxygen vacancy formation and refilling. Much of this activity has focused on the understanding of the electronic and structural properties of defective CeO(2) with and without doping, and comprehending the determining factor for oxygen vacancy formation and the rule to tune the formation energy by doping has constituted a central issue in material chemistry related to ceria. However, the calculation on electronic structures and the corresponding relaxation patterns in defective CeO(2-x) oxides remains at present a challenge in the DFT framework. A pragmatic approach based on density functional theory with the inclusion of on-site Coulomb correction, i.e. the so-called DFT + U technique, has been extensively applied in the majority of recent theoretical investigations. Firstly, we review briefly the latest electronic structure calculations of defective CeO(2)(111), focusing on the phenomenon of multiple configurations of the localized 4f electrons, as well as the discussions of its formation mechanism and the catalytic role in activating the O(2) molecule. Secondly, aiming at shedding light on the doping effect on tuning the oxygen vacancy formation in ceria-based solid solutions, we summarize the recent theoretical results of Ce(1-x)Zr(x)O(2) solid solutions in terms of the effect of dopant concentrations and crystal phases. A general model on O vacancy formation is also discussed; it consists of electrostatic and structural relaxation terms, and the vital role of the later is emphasized. Particularly, we discuss the crucial role of the localized structural relaxation patterns in determining the superb oxygen storage capacity in kappa-phase Ce(1-x)Zr(1-x)O(2). Thirdly, we briefly discuss some interesting findings for the oxygen vacancy formation in pure ceria nanoparticles (NPs) uncovered by DFT calculations and compare those with the bulk or extended surfaces of ceria as well as different particle sizes, emphasizing the role of the electrostatic field in determining the O vacancy formation.

Face classification is a technique used in Biometrics to help distinguish between facial images. However, this technique has been applicable on human face images only. Online virtual worlds such as Second Life, Sims Online, etc. are gaining popularity over the Internet. They require human users to create a digital persona of oneself, known as an "avatar". Several avatars are designed to resemble human users. With crime being reported in virtual worlds, computer-generated avatar faces being created from human faces and human-resembling humanoids being designed, there is a need to distinguish between natural and artificial faces. Our work applies two new face classification techniques on grayscale, facial images of humans and avatars to tell them apart. (1) Uniform Local Directional Pattern (ULDP) utilizes the uniform patterns from Local Directional Pattern (LDP) (2) Wavelet Uniform Local Directional Pattern (WULDP) applies the ULDP technique on the wavelet transform of an image. Extensive experiments conducted on five different face image datasets (Caltech, FERET for human faces and Entropia, Second Life, Evolver for avatar faces) achieve baseline average classification accuracies of 98.55% using ULDP and 89.55% using WULDP respectively.