Search

Cellular network fingerprint localization technology utilizes signal feature analysis for location estimation, providing a crucial technological pathway to enhance the accuracy of location-based services (LBS) in Internet of Things (IoT) applications. However, this technology faces new challenges. Firstly, to address errors in sub-model outputs that traditional hybrid prediction models cannot effectively identify, we designed a novel hybrid prediction model. This model combines a one-dimensional convolutional neural network and a fully connected neural network in parallel, integrating a newly designed filtering process to eliminate most errors in sub-model outputs and reduce subsequent computational burden. Secondly, considering the limitation of fingerprint localization technology in providing localization error information concurrently with predicted locations, we proposed a localization error estimation method. This method offers uncertainty metrics with each predicted location to provide relevant uncertainty measures. Lastly, addressing the current lack of credibility analysis for fingerprint predicted locations, we devised a credibility assessment method aimed at enhancing the reliability of localization results by providing comprehensive information. Carefully selected evaluations in complex urban environments validate the effectiveness of the proposed localization technology. Compared to existing technologies, it demonstrates superior performance with a median localization error of 8.53 m and an average localization error of 13.36 m. In the future, the technology is expected to play a key role in LBS for the IoT, improving the accuracy and reliability of the system. [ABSTRACT FROM AUTHOR]

• Cell counts of the same, Diamond™ dye (DD) stained, datasets differed when counted by different people. • Shedder categorisation using cell counting is area size (portion of a print used for counting) and counting method dependent. • Cell numbers in each shedder class are dependent on the type and number of persons tested. • Cell deposit differences exist between and within people. • DD-stained worn gloves may provide a suitable alternative for bare hands in TPPR investigations. Touch DNA is one of the most common types of biological material collected during criminal investigations. Diamond™ Nucleic Acid Dye (DD) has been shown to aid in touch sample visualisation and target sampling. It has also been used as a method of shedder categorisation that is cheaper and quicker than DNA methods. However, the DD method routinely involves manual cell counting, which can result in intra and inter-person variability similar to other manual techniques used in forensic science, for example, fingerprint identification. Additionally, DD based shedder categorisation involves counting cells in a portion of the touch deposit to extrapolate an individual's shedder status, and the sampling effect of such estimations is currently unknown. The present study tested different data analysis aspects of the DD method, including counting variability within and between people, shedder classification differences based on different counting methods (entire thumbprint, sub-section of a print with most cells, sub-section of a print deemed most representative of the entire thumbprint, and random sections), the use of ImageJ software to semi-automate counting and the use and extension of the DD method for investigating DNA Transfer, Persistence, Prevalence and Recovery (DNA-TPPR). The results of this study show that there are meaningful differences observed during counting processes both between and within people. These differences tended to increase as the factor of time, or the duration of counting, rather than the complexity of cell deposits being assessed. Investment in cell counting software that eliminates personal factors, such as boredom fatigue, can remedy most of these issues, however, will require optimisation, such as fibre recognition. Shedder testing was shown to be affected by the choice of sampling and categorisation methods, and suggested that using an entire finger or larger section size can provide increased precision. Finally, inverted worn gloves stained with DD may provide an acceptable alternative for hands in DNA-TPPR investigations, providing an interesting alternative for future research. [ABSTRACT FROM AUTHOR]

• A national fingerprint collaborative exercise carried out in 2022-23 within the UK. • Exercise incorporated fingermark visualisation and fingerprint comparisons aspects. • Participants request to process items containing blood contaminated fingermarks. • Experiences shared and results discussed amongst fingerprint practitioners at an in-person workshop. • Evidence of good practice highlighted and areas for improvement suggested. In 2022–2023 the Defence Science and Technology Laboratory (Dstl) designed and conducted a two-part fingerprint Collaborative Exercise (CE). The CE focussed on fingermarks contaminated in blood as they provide a complex scenario for forensic units and it involved both fingermark visualisation and comparison elements. Participants were requested to treat the exercise as a major crime submission following internal protocols and were invited to provide comments relating to how the marks were deposited (if appropriate). Overall, forensic units performed well within both parts of the exercise. The exercise highlighted the importance of conducting sequential fingermark visualisation techniques, utilising additional lighting techniques and maintaining detailed notes throughout the fingerprint examination process. The outputs from the CE provided opportunities for forensic units to learn from one another, raise the level of understanding of blood contaminated fingermarks and identify areas of improvement which can be incorporated into internal processes. [ABSTRACT FROM AUTHOR]

[Display omitted] • Effect of high concentration commercial corrosive substances on fingermark recovery studied. • Range of substrates, corrosive substances and age of marks extended beyond previous research. • Fingermarks still developed after exposure to concentrated (18 M) sulfuric acid. • Alkali solutions more detrimental to fingermarks than acids of equivalent concentration. • Powder suspensions found to be the most effective processes for fingermark recovery. Fingermark recovery plays a crucial role in investigating corrosive substance attacks, which are becoming increasingly common. Building upon previous research, this study aimed to identify effective visualization processes for recovering fingermarks from diverse substrates exposed to wide range of commercially available corrosive materials. The study investigated glass, PVC and HDPE substrates with fingermarks deposited 1 day and 2 weeks before exposure to the corrosive substance, and used commercially available substances at concentrations higher than any previous study. It was found that fingermarks could still be recovered from all substrates studied after exposure to most of the corrosive substances, although in general exposure to corrosive substances was detrimental to the quality of marks recovered. The most detrimental corrosive substances were found to be those based on concentrated sulfuric acid. Black and white powder suspensions were the most effective of all processes used in this study, with the highest recovery rates observed from the glass substrate. Age of mark was not found to have a significant effect on recovery rates. Overall the results show that fingermarks may survive exposure to even the most concentrated acids used in this study and provide the initial basis for guidance on processes that could be used on materials used in corrosive substance attacks. [ABSTRACT FROM AUTHOR]

Traditional magnetic resonance imaging (MRI) is qualitative imaging, and doctors need to rely on experience to diagnose diseases, which cannot meet the current needs of precision medicine. As a new quantitative magnetic resonance imaging technology, magnetic resonance fingerprint imaging can obtain a variety of human tissue parameters at the same time through a data acquisition, which greatly improves the imaging speed and improves the impact of noise on image quality. Several pattern matching algorithms are compared, including direct matching method, Bloch response iterative projection method, covering tree and approximate nearest neighbor search method, and improved methods. Absolute error image, mean absolute error (MAE), normalized root means square error (RMSE) and running time are counted in the experimental results. The results show that the improved method is better than the traditional method, which can greatly improve the quality of MR fingerprint multi-parameter images (T1, T2, B0, PD), and make the running time within an acceptable range. In addition, the improved algorithm is insensitive to random additive noise. [ABSTRACT FROM AUTHOR]

The identification of suspects based on partial and smudged fingerprints, commonly referred to as fingermarks or latent finger-prints, presents a significant challenge in the field of fingerprint recognition. While fixed-length embeddings have shown effectiveness in recognizing rolled and slap fingerprints, the methods for matching latent fingerprints have primarily centered around local minutiae-based embeddings, failing to fully exploit global representations for matching purposes. Consequently, enhancing latent fingerprints becomes critical to ensuring robust identification for forensic investigations. Current approaches often prioritize restoring ridge patterns, overlooking the fine minutiae details crucial for accurate fingerprint recognition. To tackle this, we propose a novel approach that utilizes Generative Adversarial Networks (GANs) to redefine Latent Fingerprint Enhancement (LFE) through a structured approach to fingerprint generation. By directly optimizing minutiae information during the generation process, the model produces enhanced latent fingerprints exhibiting exceptional fidelity to ground truth instances. This leads to a significant improvement in identification performance. Our framework integrates minutiae locations and orientation fields, ensuring the preservation of both local and structural fingerprint features. Extensive evaluations conducted on two publicly available datasets demonstrate our method's dominance over existing state-of-the-art techniques, highlighting its potential to significantly enhance latent fingerprint recognition accuracy in forensic applications. [ABSTRACT FROM AUTHOR]

Adverse drug reactions have, in some cases, devastating effects on the population, so it is very important to develop decision support systems based on artificial intelligence to predict such situations having a preventive effect with an important positive impact on population health. This research aims to study the influence of features, which are different characteristics used to represent chemical drug compounds, on the prediction of adverse reactions. A chemical fingerprint is defined as being a distinctive characteristic or pattern that indicates the presence of a certain molecule and a molecular descriptor that is a mathematical representation of a molecule's properties, which describes the chemical information of the molecule. Based on a comprehensive study of the state-of-the-art artificial intelligence algorithms such as Random Forest (RF), Gradient Boosted Trees (GBT), Feedforward Neural Network (FFN) and Convolutional Neural Network (CNN) are chosen, with which prediction models were created with different approaches to features such as the use of fingerprints, the use of molecular descriptors or the combination of fingerprints and molecular descriptors in order to obtain improved results. The algorithms specified above were selected since they are powerful and can capture complex patterns along the given features. The experimental evaluations consisted of identifying four effective fingerprints, namely: Pattern, Torsion, Atom Pair and Morgan, creating a machine learning model for each fingerprint type and using them as input data features, observing the results, after which two types of molecular descriptors were chosen, namely, molecular weight and number of valence electrons, we combined them as a feature of the input data, created machine learning models which we trained and observed the results, then, the last experiment, consisted of combining the molecular descriptors and the fingerprints, creating the machine learning models, training and observing the results. According to the experimental evaluation results RF and GBT had better results, with significant differences of even 20% in accuracy in certain cases, such as using molecular descriptors for Hepatobiliary disorders, than FNN and CNN. Another conclusion that can be formulated is the fact that the fingerprints had the greatest impact on the performance increase of the models, so the use of fingerprints and molecular descriptors together as features gave the best results among all the performed tests. [ABSTRACT FROM AUTHOR]

Contactless fingerprint authentication has gained popularity as a field of research in biometrics in recent years. Unlike traditional fingerprint recognition systems that require direct contact of the person's finger with the sensor, contactless fingerprint systems offer several advantages, among them ease of capture and cost-effectiveness. Despite the progress made in this field, poor contrast, background noise, and limited image information continue to pose difficulties for fingerprint recognition in contactless environments. Furthermore, the number of images in published fingerprint biometric datasets for each person is restricted, and there is insufficient data for efficient training. Nevertheless, Convolutional Neural Network (CNN) architectures have been widely used, necessitating large databases. To address these issues, this paper introduces a Siamese network designed for the purpose of identity Verification, using the contactless thumb fingerprint modality to enhance recognition results. The Siamese network is able to extract pertinent features from noisy images with low contrast and limited information, even if they have limited information and low contrast. Additionally, this work proposes the use of the contactless thumb fingerprint modality instead of the contactless index fingerprint modality, which is more commonly used in related works. Consequently, the Mobile FingerPrint (MFP) dataset is introduced and constructed for evaluation. Experimental results demonstrate the efficiency of the proposed method, achieving an accuracy of 98.68% for thumb fingerprint Verification. [ABSTRACT FROM AUTHOR]

J Med Internet Res 2022;24(11):e38650 doi:10.2196/38650 [ABSTRACT FROM AUTHOR]

Biometric identification systems are increasingly important today compared to traditional recognition/classification systems. Electromyography (EMG) signals and person identification/classification systems are preferred for high-security systems as they include physiological and behavioural movements. This study investigates biometric EMG signals based on convolutional neural networks (CNNs) and personal identification/classification systems. Bioelectric signals were recorded at six different wrist movements from five volunteer participants with a four-channel EMG device. To determine the spectrum characteristics of EMG signals, the frequency subbands of the signals were found using the discrete wavelet transform (DWT), empirical wavelet transform (EWT), and empirical mode decomposition (EMD) methods. In addition, statistical methods are used to improve the effectiveness of the feature vector. The CNN model was used to define or classify people. The performance of the developed system was evaluated using Accuracy, Precision, Sensitivity, Fscore parameters. As a result, a classification success of 95.66 % was achieved with the developed EMD-CNN method, 94.10 % with the DWT-CNN method, and 93.33 % with the EWT-CNN method. The artificial intelligence model presented in this study explains the effectiveness of EMG signals in person recognition or classification as a biometric identification system. Furthermore, the developed model shows promise for the development and design of future biometric recognition systems. [ABSTRACT FROM AUTHOR]

• Hand and finger determination will expedite identification on isolated fingerprints. • Determining the hand in fingerprints with a two or one-delta pattern is possible. • Hand determination from isolated fingerprints is possible. The individuality and permanence of fingerprints make of them a very useful feature in the identification of individuals. There are now automated computer programmes that allow a quick comparison between a fingermark and a database. However, in order to assess the strength of evidence in fingerprint identification, complementary information on the frequencies of the different morphological features of the dermopapillary ridges is required. This idea is used in this work as a starting point to evaluate the frequencies of the parameters used in the determination of the hand and finger in a large sample of 2600 fingerprints taken from 134 male and 127 female Spanish population. Based on these fingerprints, the frequencies of different categories of the following parameters were obtained: type of pattern, slope of the apex ridge, subtype of two-delta pattern, ridge tracing, major angle, major ridge count, bisector, rotation of the central ridge, assimilation to loops and slant. Moreover, the results have shown that these characters are useful for the determination of the hand in whorls (two-delta pattern) and loops (one-delta pattern), but not for the determination of the finger. The most useful and classificatory parameter when determining the hand of origin of a two-delta fingerprint is the slope of the apex ridge, and for the one-delta pattern, knowing the location of the delta allows the correct estimation of the hand of a fingerprint in more than 93% of the cases. The data presented in this paper are novel and can be used by latent print examiners to improve the statistical basis of their decisions in reaching conclusions. [ABSTRACT FROM AUTHOR]

Biometric technology, founded on recognizing distinctive individual features, enhances identification and automation, yielding personal and institutional benefits. This literature review employs the Kitchenham and Charters methodology to analyze recent perspectives on biometric technology in educational time and attendance. The analysis identifies benefits, such as preventing impersonation and enhancing data security, but emphasizes concerns about biometric alterations and hardware/software failures hindering optimal outcomes. A noteworthy trend is adopting a bimodal behavioral approach in biometric technologies in education. The study recommends research avenues focusing on privacy safeguards, biometric system robustness, and technological resilience. Beyond mere identification, a shift towards quantitative and descriptive methods is proposed, guiding future research. The results underscore the importance of a multimodal approach and stringent security measures to address biometric system vulnerabilities. Future exploration should consider integrating emerging technologies to enhance system resilience and data protection, providing a comprehensive guide for refining and expanding biometric technology applications in education. [ABSTRACT FROM AUTHOR]

Nowadays, authentication remains a crucial and primordial process in securing Information Technology (IT) systems and infrastructures, while keeping users' data more secure and privacy-preserving. Indeed, finding a distinctive characteristic(s) for each individual is done through personal identification and verification. Therefore, techniques such as Two-Factor Authentication (2FA) have been widely used to increase the security level in the authentication processes; where passwords are combined with secret codes or Unimodal Biometrics (UB). In this research study, we consider the case of 2FA based on UB, where the users are mainly asked each time to choose the suited biometric trait for their current authentication process. The thing that can obviously lead to a degradation of the User Experience (UX) and increases the Authentication Time (AT). To overcome these shortcomings, we propose in this article an "Autonomous Unimodal Biometric Authentication System", abbreviated to "AUBAS", which is modeled and developed on the basis of discrete-time Markov chains. It is a lightweight and seamless system that allows authenticating users according to two strategies, namely "likelihood" and "weighting". Eventually, we take a case study and evaluate the system performance with the necessary simulations. Consequently, the implementation of AUBAS allows for improving the UX of an IT system and can be further applied in any environment or process relying on unimodal biometric authentication, such as cloud or mobile computing. [ABSTRACT FROM AUTHOR]

• Mueller polarimetry develops latent fingerprints and removes their backgrounds. • The procedure is non-destructive, so it preserves evidence for further analysis. • The method is flexible and robust against measurement errors. • The proposed instrument is inexpensive which makes it accessible for any crime lab. • The Mueller polarimeter is portable so it is suitable for crime scene work. In this article we have explored the use of Mueller polarimetry for the simultaneous and non-destructive latent fingerprint development and background elimination. This proposal not only expands previously reported uses of the polarization state of light for fingerprint development, but offers some advantages. Samples of a few donors, taken on different heavily back grounded substrates, were measured and processed. Several strategies for background removal were applied. The results are very promising, since the background was successfully erased in all cases and the quality of the developed fingerprints was excellent for most substrates and donors. Considering that the method is non-destructive and that the proposed measurement instrument is inexpensive and portable the method could be tried in real cases with no risk for the evidence. In addition, the success of the procedure for background removal raises the possibility of exploring if it would work for the enhancement of fingerprints developed with other methods. [ABSTRACT FROM AUTHOR]

With the emergence of deep learning, Convolutional Neural Network (CNN) models have been proposed to advance the progress of various applications, including face recognition, object detection, pattern recognition, and number plate recognition. The utilization of CNNs in these areas has considerably improved security and surveillance capabilities by providing automated recognition solutions, such as traffic surveillance, access control devices, biometric security systems, and attendance systems. However, there is still room for improvement in this field. This paper discusses several classic CNN models, such as LeNet-5, AlexNet, VGGNet, GoogLeNet, and ResNet, as well as lightweight models for mobile-based applications, such as MobileNet, ShuffleNet, and EfficientNet. Additionally, deep CNN-based face recognition models, such as DeepFace, DeepID, FaceNet, and SphereFace, are explored, along with their architectural characteristics, advantages, disadvantages, and recognition accuracy. The results indicate that many scholars are researching lightweight face recognition, but applying it to mobile devices is impractical due to high computational costs. Furthermore, noise label learning is not robust in actual scenarios, and unlabeled face learning is expensive in manual labeling. Finally, this paper concludes with a discussion of the current problems faced by face recognition technology and its potential future directions for development. [ABSTRACT FROM AUTHOR]

• Fingermark enhancement processes require performance monitoring. • Existing methods lack a reliable, consistent and quantifiable approach. • Previous attempts require specialist equipment and lack broader substrate applicability. • This work presents confirmatory tests for amino acid, sebaceous and cyanoacrylate based reagents. • Tests have a proven shelf-life longevity in excess of twelve months. Validation is particularly important in forensic science. Each process utilised by a forensic laboratory will undergo validation to ensure it is functioning as expected. Some pieces of equipment (hardware) will be simple to validate, and this will be conducted regularly. However, many commonly employed fingermark enhancement reagents, such as ninhydrin and 1,2-Indandione, do not have clearly defined and robust validation parameters. To that end, artificial fingermark solutions, particularly amino acid solutions, have previously been investigated as a controlled validation method, via deposition onto a substrate. Whilst amino acid reagents are important, there is also a wider group of latent fingermark constituents (targets) that require the use of other enhancement reagents within the forensic laboratory. The work presented herein seeks to expand on the concept of amino acid printing, by exploring the possibilities of depositing a sebaceous solution matrix, thereby facilitating the testing of reagents such as Oil-Red-O and Physical Developer. In addition, we present a method that extends capabilities beyond fingermark enhancement reagents for porous substrates, by developing a methodology that enables process validation of the cyanoacrylate fuming technique to be easily facilitated. These simple and effective solutions have the capacity to serve as a crucial process validation check within the laboratory validation workflow. [ABSTRACT FROM AUTHOR]

Mobile crowdsensing leverages the ubiquitous sensors of smart devices to facilitate various sensing applications. Users who participate in contributing data usually get rewards from the task requester, while there is a potential risk that someone would preempt the task and provide a forged sensing report for seeking revenue with minimal effort. Thus, trust assessment is essential to identify those irregular sensing reports. The existing methods mainly consider users' reputations and estimate the trustworthiness upon the difference from the aggregated result. However, they still face a severe problem when a majority of reports are invalid or low-quality caused by the repeated submission, e.g., a user can switch multiple accounts on a single device to repeatedly submit forged reports. To tackle this problem, we design a trust assessment scheme with an enhanced device fingerprinting algorithm. Briefly, to reduce the influence of the repeated sensing reports, we first compute their unique fingerprints derived from the intrinsic characteristics of sensors and assign an initial trust weight for each report. Then, to improve the accuracy of the assessment, we further compute the similarity of the reports to obtain their final trust values. Extensive evaluations are conducted to justify the effectiveness of our proposed design. • We study the case when most of the reports are invalid due to repeated submission. • The intrinsic characteristics of device sensors are extracted for trust assessment. • Simulation and experiment results justify the effectiveness of the proposed design. [ABSTRACT FROM AUTHOR]

Natural artificial facial expressions are necessary for the development of interactive virtual communication technologies. One effort that can be made to form compound facial expressions is a combination of several basic emotional classes. Identification of the dominant basic emotional classes in sentences can be done using a multilabel classification process. In this study, we combined the Multichannel Convolution Neural Network model and multilabel techniques to improve classification performance in identifying dominant basic emotional classes. Compound facial expressions are formed by combining these dominant emotional classes. The data used in the classification process is the text of social media posts that have been labeled by analysis from linguists into six categories of emotions: fear, anger, disgust, sadness, happiness, and surprise. The results of the labeling process were validated by psychologists. The experimental results for the 8:2 data portion showed that the CNN Multichannel model with kernel size = [4, 5, 6] and max-pooling layers produced the average value of the highest accuracy of 96% and F1-score of 86% compared to the use of the Simple bidirectional LSTM model with 32 LSTM units, CNN Single channel and other variations of CNN Multichannel. The experimental results show that our proposed method has better performance than some of the CNN models used in previous studies. [ABSTRACT FROM AUTHOR]

A new lightweight traffic sign detection algorithm based on YOLOv4 has been proposed to address the time-consuming issue of existing algorithms when training more parameters. The algorithm involves several key improvements to YOLOv4's backbone network and feature pyramid. Specifically, MobileNetv3 is used as the backbone feature extraction network, replacing standard convolution with depthwise separable convolution to reduce the number of parameters and computation in both the backbone network and feature pyramid. Furthermore, two SPP modules are added to the feature pyramid part to improve the detection accuracy of the model. The MobileNetv3 network is also improved by dropping the modules after layer 17 and performing PW (PointWise) convolution operations on feature layers to convert their dimensions and connect them with the detection layer. Finally, the a priori frame is initialized by applying the K-means++ algorithm to further enhance the detection accuracy of the algorithm. Experimental results on the CCTSDB dataset indicate that the proposed algorithm achieves a 1.7% increase in mAP (mean average precision), reduces the parameter amount by 197M, and increases detection speed by 25% compared to YOLOv4. These results demonstrate that the improved lightweight algorithm performs exceptionally well in traffic sign detection. [ABSTRACT FROM AUTHOR]

The high-altitude and mobility characteristics of unmanned aerial vehicles (UAVs) have made them a key element of new radio systems, particularly because they can exceed the limits of terrestrial networks. However, at high altitudes, UAVs can be significantly affected by intercell interference at a high line-of-sight probability. To mitigate this drawback, we propose an algorithm that selects the optimal beam to reduce interference and maximize transmission efficiency. The proposed algorithm comprises two steps: constructing a user-location-based fingerprint database according to the user types presented herein and cooperative beam selection. Simulations were conducted using cellular cooperative downlink systems for analyzing the performance of the proposed method, and the signal-to-interference-plus-noise cumulative distribution function and spectral efficiency cumulative distribution function were used as performance analysis indicators. Simulation results showed that the proposed algorithm could reduce the effect of interference and increase the performance of the desired signal. Moreover, the algorithm could efficiently reduce overheads and system cost by reducing the amount of resources required for information exchange. [ABSTRACT FROM AUTHOR]

Hand tracking and identification through visual means pose a challenging problem. To simplify the identification of hand gestures, some systems have incorporated position markers or colored bands, which are not ideal for controlling robots due to their inconvenience. The motion recognition problem can be solved by combining object identification, recognition, and tracking using image processing techniques. A wide variety of target detection and recognition image processing methods are available. This paper proposes novel CNN-based methods to create a user-free hand gesture detection system. The use of synthetic techniques is recommended to improve recognition accuracy. The proposed method offers several advantages over existing methods, including higher accuracy and real-time hand gesture recognition suitable for sign language recognition and human-computer interaction. The CNN automatically extracts high-level characteristics from the source picture, and the SVM is used to classify these features. This study employed a CNN to automatically extract traits from raw EMG images, which is different from conventional feature extractors. The SVM classifier then determines which hand gestures are being made. Our tests demonstrate that the proposed strategy achieves superior accuracy compared to using only CNN. [ABSTRACT FROM AUTHOR]

Purpose It is important to comprehend how the molecular mechanisms shift when gastric cancer in its early stages (GC). We employed integrative bioinformatics approaches to locate various biological signalling pathways and molecular fingerprints to comprehend the pathophysiology of the GC. To facilitate the discovery of their possible biomarkers, a rapid diagnostic may be made, which leads to an improved diagnosis and improves the patient's prognosis. Methods Through protein--protein interaction networks, functional differentially expressed genes (DEGs), and pathway enrichment studies, we examined the gene expression profiles of individuals with chronic atrophic gastritis and GC. Results A total of 17 DEGs comprising 8 upregulated and 9 down-regulated genes were identified from the microarray dataset from biopsies with chronic atrophic gastritis and GC. These DEGs were primarily enriched for CDK regulation of DNA replication and mitotic M-M/G1 phase pathways, according to KEGG analysis (p > 0.05). We discovered two hub genes, MCM7 and CDC6, in the protein--protein interaction network we obtained for the 17 DEGs (expanded with increased maximum interaction with 110 nodes and 2103 edges). MCM7 was discovered to be up-regulated in GC tissues following confirmation using the GEPIA and Human Protein Atlas databases. Conclusion The elevated expression of MCM7 in both chronic atrophic gastritis and GC, as shown by our comprehensive investigation, suggests that this protein may serve as a promising biomarker for the early detection of GC. [ABSTRACT FROM AUTHOR]

Advancements in remote medical technologies and smart devices have led to expectations of contactless rehabilitation. Conventionally, rehabilitation requires clinicians to perform routine muscle function assessments with patients. However, assessment results are difficult to cross‐reference owing to the lack of a gold standard. Thus, the application of remote smart rehabilitation systems is significantly hindered. This study analyzes the factors affecting the real‐time evaluation of muscle function based on biometric sensor data so that we can provide a basis for a remote system. We acquired real clinical stroke patient data to identify the meaningful features associated with normal and abnormal musculature. We provide a system based on these emerging features that assesses muscle functionality in real time via streamed biometric signal data. A system view based on the amount of data, data processing speed, and feature proportions is provided to support the production of a rudimentary remote smart rehabilitation system. [ABSTRACT FROM AUTHOR]

With the development of science and technology, aspherical surface has become an indispensable component in optical system. However, the aspheric surface has more complex optical properties than the spherical surface, so it restricts the rapid development of aspheric surface. In this paper, the main problem of aspherical machining is analyzed, which is"Trajectory control", and the unique optical properties of aspherical surface are analyzed, a trajectory control algorithm based on genetic-BP neural network is proposed to solve the trajectory control in aspheric forming. The method is proved to be feasible and effective. [ABSTRACT FROM AUTHOR]

Atherosclerotic plaque rupture in carotid arteries is a major cause of cerebrovascular events. Plaque rupture is the mechanical failure of the heterogeneous fibrous plaque tissue. Local characterization of the tissue's failure properties and the collagen architecture are of great importance to have insights in plaque rupture for clinical event prevention. Previous studies were limited to average rupture properties and global structural characterization, and did not provide the necessary local information. In this study, we assessed the local collagen architecture and failure properties of fibrous plaque tissue, by analyzing 30 tissue strips from 18 carotid plaques. Our study framework entailed second harmonic generation imaging for local collagen orientation and dispersion, and uniaxial tensile testing and digital image correlation for local tissue mechanics. The results showed that 87% of the imaged locations had collagen orientation close to the circumferential direction (0°) of the artery, and substantial dispersion locally. All regions combined, median [Q1:Q3] of the predominant angle measurements was -2° [-16°:16°]. The stretch ratio measurements clearly demonstrated a nonuniform stretch ratio distribution in the tissue under uniaxial loading. The rupture initiation regions had significantly higher stretch ratios (1.26 [1.15-1.40]) than the tissue average stretch ratio (1.11 [1.10-1.16]). No significant difference in collagen direction and dispersion was identified between the rupture regions and the rest of the tissue. The presented study forms an initial step towards gaining better insights into the characterization of local structural and mechanical fingerprints of fibrous plaque tissue in order to aid improved assessment of plaque rupture risk. Plaque rupture risk assessment, critical to prevent cardiovascular events, requires knowledge on local failure properties and structure of collagenous plaque tissue. Our current knowledge is unfortunately limited to tissue's overall ultimate failure properties with scarce information on collagen architecture. In this study, local failure properties and collagen architecture of fibrous plaque tissue were obtained. We found predominant circumferential alignment of collagen fibers with substantial local dispersion. The tissue showed nonuniform stretch distribution under uniaxial tensile loading, with high stretches at rupture spots. This study highlights the significance of local mechanical and structural assessment for better insights into plaque rupture and the potential use of local stretches as risk marker for plaque rupture for patient-specific clinical applications. [Display omitted] [ABSTRACT FROM AUTHOR]

A low-cost, readily accessible phosphorescent powder is herein presented capable of negating background substrate interference after charging with white light. [Display omitted] • Doped strontium aluminate materials were explored as fingerprint powders. • Blue, green, and aqua phosphorescence persisting for > 30 s was seen. • The powders can be charged using white light. • The powders are commercially available and inexpensive to purchase. An inexpensive, commercially available doped strontium aluminate phosphor with long-lived afterglow has been prepared and assessed in the role of a luminescent fingerprint dusting powder. Blue, green, and aqua phosphorescence persisting for ca. 30 s was obtainable from treated fingermarks after charging the powders with the white light (400–700 nm) setting of a forensic light source. Imaging the phosphorescent afterglow enabled the elimination of background emissions encountered during latent fingermark examination. This was demonstrated by visualising fingermarks on substrates that possess inbuilt fluorescent security features and highly patterned substrate backgrounds, without any need for bespoke scientific equipment. [ABSTRACT FROM AUTHOR]

With rising concerns about data theft and stricter security rules in many countries, biometric technology is becoming increasingly integral to our everyday lives. Given the severe limits of current single-modal biometric systems, it is no surprise that multimodal biometric approaches are experiencing a surge in popularity. Based on these findings, this work provides a novel multimodal biometric person identification system that combines iris, face, and palmprint biometric modalities for human recognition via the use of deep learning algorithms. The network relies on a convolutional neural network (CNN) to extract features, and a SoftMax and Tanh classifier to label images. The Adam and Adadelta optimisation technique are utilised to construct the CNN model, and the categorical cross-entropy loss function was implemented. As a result, the functional and evaluation levels were fused together. Several tests were conducted on the PolyU-IITD, PolyU Cross-Spectral Iris Image, and Tufts Face datasets to empirically evaluate the performance of the proposed system. In a biometric identification system, employing three biometric characteristics was shown to be superior to using one or two biometric features. Furthermore, the findings demonstrate that the proposed method achieves 100% accuracy, much better than existing state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

The purpose of the presented research was to evaluate the performance of the prepared biometric algorithms and obtain information on the opinions and preferences of their users in bank branches. The study aimed to determine users' attitudes towards particular modalities and preferences on how to use biometrics after the bank customers had practical experience with the operation of the prototype solutions. The research results allow a better understanding of users' needs and expectations, which can help to increase the knowledge of the need for biometric solutions in banking. Making appropriate changes based on the research results can help improve user satisfaction. The experiments included the collection of biometric sample collection and verification processes conducted by bank customers assisted by employees. The study collected biometric samples and questionnaires from 365 people, and 179 complete verifications were conducted. About 95% of those surveyed expressed acceptance of the implemented solutions for handling transactions using biometrics. The HandVein-FaceImage (hand vein scanning and face recognition) modality pair was rated highest in 70% of cases in the studied population. [ABSTRACT FROM AUTHOR]

A virtual coaching system is key for monitoring the performance during exercise regimes and rendering timely feedback to avoid potential physical harm. Ideally, the core technologies behind such a system imply exercise recognition/assessment for performance evaluation and subject identification for retrieval/update of personal profiles. In this work, we develop techniques based on learning models for inferences regarding exercise recognition, prediction of the subject's biometrics, exercise evaluation, and simultaneous recognition and evaluation. We address four indoor exercises: squats, push-ups, shoulder-press, and lunges. Different sorts of exercise evaluation are provided: (1) the overall-binary assessment, with a detailed assessment rendering the exact mistakes performed, and (2) the score-based evaluation giving a metric in the range of 0-10 for each aspect of the exercise. Thermal and RGB imagery are examined with various descriptors, such as raw images, Gait Energy Image (GEI), and skeletal pose, with a comparison between their performance. The GEI descriptor extracted from RGB imagery is introduced for exercise recognition and evaluation tasks achieving remarkable results. The GEI descriptor is further utilized for subject recognition and verification from the performed exercise showing the ability of subject discrimination from the way of exercising. Consequently, we introduce new kinds of biometrics based on indoor exercising inspired by the notion that individuals are recognizable from the fingerprint motion dynamics in exercise regimes. To the best of our knowledge, our work is the first to use GEI in such directions and show its remarkable effectiveness. [ABSTRACT FROM AUTHOR]

With the development of IoT and artificial intelligence, location-based services are getting more and more attention. For solving the current problem that indoor positioning error is large and generalization is poor, this paper proposes a Model Stacking Algorithm for Indoor Positioning System using WiFi fingerprinting. Firstly, we adopt a model stacking method based on Bayesian optimization to predict the location of indoor targets to improve indoor localization accuracy and model generalization. Secondly, Taking the predicted position based on model stacking as the observation value of particle filter, collaborative particle filter localization based on model stacking algorithm is realized. The experimental results show that the algorithm can control the position error within 2m, which is superior to KNN, GBDT, Xgboost, LightGBM, RF. The location accuracy of the fusion particle filter algorithm is improved by 31%, and the predicted trajectory is close to the real trajectory. The algorithm can also adapt to the application scenarios with fewer wireless access points. [ABSTRACT FROM AUTHOR]

The Graph Partitioning Problem (GPP) is a classical combinatorial optimization problem that has been extensively researched. In recent years, many methods for solving the GPP have been proposed, which can be divided into direct partitioning approaches and iterative improvement approaches. Direct partitioning approaches directly give a complete partition of the input graph. Iterative improvement approaches require further adjustment of feasible solutions based on the result of direct partitioning approaches in order to obtain a better performance. In this paper, we summarize the recent trends in algorithms and applications for GPP. In addition, we propose a graph partitioning algorithm based on semi-supervised learning in combination with graph filtering methods. [ABSTRACT FROM AUTHOR]

• A national collaborative exercise carried out in 2020 within the UK. • The exercise investigated fingermark visualisation on wrapping paper. • Variation in the approach to this exercise was expected due to its complex nature. • Key learning points and lessons learnt were identified. In 2020, the Defence Science and Technology Laboratory organised and ran what is believed to be the first UK national collaborative exercise of its kind in the field of fingermark visualisation, on behalf of the Forensic Science Regulator. Laboratories were provided with a piece of wrapping paper, a challenging item for fingermark visualisation due to its semi-porous characteristics, both from a planning and processing perspective, and asked to treat it as a major crime exhibit. Due to the complexity of the substrate, variation in approach was anticipated. 23 laboratories from 21 organisations completed the exercise. In general, laboratories performed well, providing assurance to the Forensic Science Regulator regarding their ability to visualise fingermarks. Key learning points were identified around decision-making, planning and implementation of fingermark visualisation processes – all of which assist in raising the level of understanding around the likely success of fingermark visualisation. Lessons learnt, along with the overall findings, were shared and discussed in a workshop held in summer 2021. The exercise provided a useful insight into the current operational practices of participating laboratories. Areas of good practice were identified as well as the areas within the laboratories' approach that could be altered or adapted. [ABSTRACT FROM AUTHOR]

Due to the availability of technology stemming from in-depth research in this sector and the drawbacks of other identifying methods, biometrics has drawn maximum attention and established itself as the most reliable alternative for recognition in recent years. Efforts are still being made to develop a userfriendly system that is up to par with security-system requirements and yields more reliable outcomes while safeguarding assets and ensuring privacy. Human age estimation and Gender identification are both challenging endeavours. Biomarkers and methods for determining biological age and gender have been extensively researched, and each has advantages and disadvantages. Facial-image-based positioning is crucial for many applications, including safety and security systems, border control, human engagement in sophisticated ambient analytics, and biometric identification. Determining a person's age and gender is a complex study method. With the advent of deep learning, the study of face systems has been completely transformed, and estimation accuracy is a crucial parameter for evaluating algorithms and their efficacy in predicting absolute ages. The UTKFace dataset, which serves as the backbone of the face estimating system, was used to assess the method. The eyes, cheeks, nose, lips, and forehead provide the foundation of this function. AlexNet achieves a 98% accuracy rate across its lifespan of system results. [ABSTRACT FROM AUTHOR]

The aluminum fluoride (AF) addition in aluminum electrolysis process (AEP) can directly influence the current efficiency, energy consumption, and stability of the process. This paper proposes an optimization scheme for AF addition based on pruned sparse fuzzy neural network (PSFNN), aiming at providing an optimal AF addition for aluminum electrolysis cell under normal superheat degree (SD) condition. Firstly, a Gaussian mixture model (GMM) is introduced to identify SD conditions in which the operating modes of AEP are unknown. Then, PSFNN is proposed to establish the AF addition model under normal SD condition identified by GMM. Specifically, a sparse regularization term is designed in loss function of PSFNN to extract the sparse representation from nonlinear process data. A structure optimization strategy based on enhanced optimal brain surgeon (EOBS) algorithm is proposed to prune redundant neurons in the rule layer. Mini-batch gradient descent and AdaBound optimizer are then introduced to optimize the parameters of PSFNN. Finally, the performance is confirmed on the simulated Tennessee Eastman process (TEP) and real-world AEP. Experimental results demonstrate that the proposed scheme provides a satisfactory performance. • A novel pruned sparse fuzzy neural network (PSFNN) is developed to establish the AF addition model under normal SD condition. • A sparse regularization term based on Kullback–Leibler divergence is designed in loss function of PSFNN to obtain the sparse representation. • The parameters of network are learned based on mini-batch gradient descent method with AdaBound optimizer. • An enhanced optimal brain surgeon (EOBS) algorithm is proposed to obtain a compact network structure. • The experimental results demonstrated that our method achieves a satisfactory performance. [ABSTRACT FROM AUTHOR]

Using a pre-trained Convolutional Neural Network (CNN) model for a practical biometric authentication system requires specific procedures for training and performance evaluation. There are two criteria for a practical biometric system studied in this paper. First, the system's ability to handle identity theft or impersonation attacks. Second, the ability of the system to generate high authentication performance with minimal enrollment period. We propose the use of the Multiple Clip Contrast Limited Adaptive Histogram Equalization (MC-CLAHE) technique to process finger images before being trained by CNN. A pre-trained CNN model called AlexNet is used to extract features as well as classify the MC-CLAHE images. The authentication performance of the pre-trained AlexNet model has increased by a maximum of 30% when using this technique. To ensure that the pre-trained AlexNet model is evaluated based on its ability to prevent impersonation attacks, a procedure to generate the Receiver Operating Characteristics (ROC) curve is proposed. An offline procedure for training the pre-trained AlexNet model is also proposed in this paper. The purpose is to minimize the user enrollment period without compromising the authentication performance. In this paper, this procedure successfully reduces the enrollment time by up to 95% compared to using on-line training. [ABSTRACT FROM AUTHOR]

As biometric technology evolves speedily, challenges increase to implement secure and accurate systems for personal identification and authentication as the highest target to achieve. Parallel security system evolution reveals continuously that monomodal biometrics is no longer exerted due to the lack of relevance in intruder's identification. The switch to multimodality started to be more attractive for researchers as well as investigators considering their advantages. We present a tetra modal biometric base integrating hard (fingerprint, face, and iris) and genetic (DeoxyriboNucleic Acid (DNA)) characteristics building process, extracted features, and matching scores from 500 consented volunteers throughout Tunisia. This database was developed within the framework of the PBMLT project (Plateforme Biométrique Multimodale de Lutte contre le Terrorisme). A fusion of matching scores generated by fingerprint, face, and iris modalities to identity verification is carried out using the Choquet integral approach. The collected multimodal database is flexible, suggesting that if one of the modalities is excluded, the system can still ensure security using the remaining. Except for DNA, the modality often approves its forensic relevance on its monomodal subsystem. Scores@PBMLT database will open up benchmark perspectives, particularly in terms of fusion algorithms at the score level. • A new multimodal biometric score database is presented. • It includes information about Fingerprint, Face, Iris and DNA. • Experimental protocols and baseline results are provided to promote Scores@PBMLT database usage. [ABSTRACT FROM AUTHOR]

In deep learning (DL), the deep generative model is helpful for data augmentation objectives to tackle the lack of datasets that have a significant impact on learning performance. Data augmentation or synthesis is expected to solve the issue in a small/sparse database. The problem of databasing also exists in the fingerprint-based indoor localisation system. The dense offline fingerprint database must be constructed with the accuracy requirement. However, this will affect the high cost, massive laborious work, and increase the complexity of the system. Therefore, this paper proposes to address these issues by generating synthetic data via a deep generative model. The generative adversarial network (GAN) is selected to generate the synthetic fingerprint database for indoor localisation. Our database consideration consists of power-based parameters, i.e., the received signal strength indicator (RSSI) from Wi-Fi devices obtained from the actual measurement campaign. Some of the literature mainly discusses how GAN works in a vast and complex dataset. Here, we consider applying GAN in a relatively small dataset and for a simple setup. Our results show that by only using the 20 % fraction of actual RSSI data combined with the synthetic RSSI, the accuracy validation performance is slightly higher than when using all actual data usage. Moreover, in only 60% of actual data usage and in combination with 625 samples of synthetic data, the accuracy performance is improved to 0.73 (1.37 times higher than the use of all actual data, 0.53). Thus, this result proves that the challenges of offline fingerprint databases can be alleviated by data synthesis through GAN by using only a small dataset. [ABSTRACT FROM AUTHOR]

Copyright of Ciencia y Sociedad is the property of Ciencia y Sociedad and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

• Higher sensitivity of ToF-SIMS compared with standard fingerprint processes. • ToF-SIMS produced ridge detail when the process failed in a depletion series. • ToF-SIMS enhanced ridge detail when applied after failed standard processes. • Multivariate analysis, using simsMVA, was explored to simplify image generation. The processes routinely used by police forces to visualise fingermarks in casework may not provide sufficient ridge pattern quality to aid an investigation. Time of Flight-Secondary Ion Mass Spectrometry (ToF-SIMS) has been proposed as a technique to enhance fingermark recovery. The technique is currently designated a Category C process in the Fingermark Visualisation Manual (FVM) as it shows potential for effective fingermark visualisation but has not yet been fully evaluated. Here the sensitivity of ToF-SIMS on three common exhibit-type surfaces - paper, polyethylene and stainless-steel was compared to standard processes. An adapted Home Office grading scale was used to evaluate the efficacy of fingerprint development by ToF-SIMS and to provide a framework for comparison with standard processes. ToF-SIMS was shown to visualise more fingerprints than the respective standard process, for all surfaces tested. In addition, ToF-SIMS was applied after the standard processes and successfully enhanced the fingerprint detail, even when the standard process failed to visualise ridge detail. This demonstrates the benefit for incorporating it into current operational fingermark development workflows. Multivariate analysis (MVA), using simsMVA, was additionally explored as a method to simplify the data analysis and image generation process. [ABSTRACT FROM AUTHOR]

Biometrics are the body measurements and calculations related to individuals. Biometrics validation is used as a form of identification of individual. Gait recognition system is one of the most advanced technology that people have been working on for a while now that takes center stage in the field of biometrics. Compared to the other types of existing systems of biometric recognition such as fingerprint detection, iris-scanning systems etc., Gait Recognition system ensures no human intervention. This paper focuses on recognition based on a person's gait. Every person has a distinct gait pattern that is unique to every other person. To train the model CASIA-B dataset has been used. The dataset includes 124 subjects where each sample has undergone Gait Energy Image extraction. Samples with clothing and baggage have been included which changes the silhouette of the person. Therefore, the model has been trained for a wider application where people wear different type of clothing and carry-ons. A Convolutional Neural Network consisting of 8 layers has been trained which performs well on both samples of dataset and an accuracy of 95.45% was obtained on dataset not involving layers of clothing and accuracy of 91.80% was obtained for the sample with clothing and baggage. [ABSTRACT FROM AUTHOR]

The use of biometrics enables the inspection and evaluation of several aspects of a person's personality. The prefix "bio" and the suffix "metric" each contain connotations that are related to the ideas of life and measurement, respectively. A number of different computational algorithms could be used to ascertain a person's level of authorization. Evaluating these methods can be done using either a single biometric characteristic or numerous individual biometric features taken together. Both of these options are available. It has been discovered that the finger knuckle print, also known as the FKP, is a practicable biometric feature that can be utilized to this day in order to obtain a sufficient level of security and accuracy. This study aims to provide an overview of the FKP recognition-based systems that are capable of providing accurate, secure, and trustworthy authentication of a person. This review will attempt to discover the methodology, principles, data sets, algorithm filters, and other characteristics used in the development of FKP recognition systems. [ABSTRACT FROM AUTHOR]

One of the most common and successful biometric frameworks is iris recognition, which has yielded promising results in systems of access control and identity authentication. Recent systems of iris recognition have concentrated on photos taken in non-ideal circumstances. Non-ideal imaging environments involve the capture of iris pictures in motion, at a distance, or under near-infrared (NIR) or visible wavelength illumination, which introduces noise factors such as gaze deviation, lack of focus, and obstruction by hair, eyelids, lighting, specular reflections, and eyeglasses. Segmenting irises is still a challenging task for eye images taken under non-ideal circumstances for iris recognition, and is strongly dependent on pupil segmentation; to estimate a line of sight using the pupil centre method, pupil segmentation is essential. The quality of eye pictures often changes due to noise effects and the individual variance between human eyes, making segmentation of the pupil a tricky problem. To address this issue, we propose an efficacious, swift, and robust iris recognition method that is convenient for eye pictures under non-ideal conditions. In this paper, we present a novel pupil segmentation method based on a fusion of the multi-scale gray-level co-occurrence matrix (MSGLCM) and the multi-range circle Hough transform (MRRCHT) methods. Using this segmentation approach, the pupil texture region is extracted correctly and a Hough transform is then used for the outer iris region. To support the improvements in segmentation accuracy, we also apply pre- and post-processing operations. An evaluation of the proposed method is conducted in this paper using the CASIA v4.0 (Distance) and MMU V2 databases, which reflect the challenges of iris recognition from non-ideal pictures. Extensive experimental findings clarify the effectiveness of the suggested method for non-ideal iris pictures. [ABSTRACT FROM AUTHOR]

One of the most common and successful biometric frameworks is iris recognition, which has yielded promising results in systems of access control and identity authentication. Recent systems of iris recognition have concentrated on photos taken in non-ideal circumstances. Non-ideal imaging environments involve the capture of iris pictures in motion, at a distance, or under near-infrared (NIR) or visible wavelength illumination, which introduces noise factors such as gaze deviation, lack of focus, and obstruction by hair, eyelids, lighting, specular reflections, and eyeglasses. Segmenting irises is still a challenging task for eye images taken under non-ideal circumstances for iris recognition, and is strongly dependent on pupil segmentation; to estimate a line of sight using the pupil centre method, pupil segmentation is essential. The quality of eye pictures often changes due to noise effects and the individual variance between human eyes, making segmentation of the pupil a tricky problem. To address this issue, we propose an efficacious, swift, and robust iris recognition method that is convenient for eye pictures under non-ideal conditions. In this paper, we present a novel pupil segmentation method based on a fusion of the multi-scale gray-level co-occurrence matrix (MSGLCM) and the multi-range circle Hough transform (MRRCHT) methods. Using this segmentation approach, the pupil texture region is extracted correctly and a Hough transform is then used for the outer iris region. To support the improvements in segmentation accuracy, we also apply pre- and post-processing operations. An evaluation of the proposed method is conducted in this paper using the CASIA v4.0 (Distance) and MMU V2 databases, which reflect the challenges of iris recognition from non-ideal pictures. Extensive experimental findings clarify the effectiveness of the suggested method for non-ideal iris pictures. [ABSTRACT FROM AUTHOR]

The Electrocardiogram records the heart's electrical signals. It is a practice; a painless diagnostic procedure used to rapidly diagnose and monitor heart problems. The ECG is an easy, noninvasive method for diagnosing various common heart conditions. Due to its unique advantages that other humans do not share, in addition to the fact that the heart's electrical activity may be easily detected from the body's surface, security is another area of concern. On this basis, it has become apparent that there are essential steps of pre-processing to deal with data of an electrical nature, signals, and prepare them for use in Biometric systems. Since it depends on the structure and function of the heart, it can be utilized as a biometric attribute for identification and recognition. This academic paper presents a new data formatting technique to be used in biometric systems based on the two major dataset kinds, PTB and PTB-XL. [ABSTRACT FROM AUTHOR]

Fingerprint identification relies on computations and classification models based on images to identify individuals at their most basic level. For feature extraction, several image preprocessing approaches are used, and image locality bifurcations of different kinds are used for classification. For feature extraction and classification, artificial neural networks (ANNs) are proposed. ANN machine learning method and Gabor filter are introduced in this paper for feature extraction and classification respectively. Artificial Neural Networks and Gabor filtering features are used to create the feature vector. An algorithm based on the extracted features was developed to create a multiclass classifier. Special Database - NIST SD4 served as the basis for evaluation in this research. The Error matrix led to the discovery that, in terms of accuracy, the approach was superior to many traditional machine learning algorithms like Support Vector Machine, Random Forest, Decision Tree and KNN. [ABSTRACT FROM AUTHOR]