User Recognition Research Articles

Exploring the Role of Machine Learning in Diagnosing and Treating Speech Disorders: A Systematic Literature Review.

Speech disorders profoundly impact the overall quality of life by impeding social operations and hindering effective communication. This study addresses the gap in systematic reviews concerning machine learning-based assistive technology for individuals with speech disorders. The overarching purpose is to offer a comprehensive overview of the field through a Systematic Literature Review (SLR) and provide valuable insights into the landscape of ML-based solutions and related studies. The research employs a systematic approach, utilizing a Systematic Literature Review (SLR) methodology. The study extensively examines the existing literature on machine learning-based assistive technology for speech disorders. Specific attention is given to ML techniques, characteristics of exploited datasets in the training phase, speaker languages, feature extraction techniques, and the features employed by ML algorithms. This study contributes to the existing literature by systematically exploring the machine learning landscape in assistive technology for speech disorders. The originality lies in the focused investigation of ML-speech recognition for impaired speech disorder users over ten years (2014-2023). The emphasis on systematic research questions related to ML techniques, dataset characteristics, languages, feature extraction techniques, and feature sets adds a unique and comprehensive perspective to the current discourse. The systematic literature review identifies significant trends and critical studies published between 2014 and 2023. In the analysis of the 65 papers from prestigious journals, support vector machines and neural networks (CNN, DNN) were the most utilized ML technique (20%, 16.92%), with the most studied disease being Dysarthria (35/65, 54% studies). Furthermore, an upsurge in using neural network-based architectures, mainly CNN and DNN, was observed after 2018. Almost half of the included studies were published between 2021 and 2022).

Touchscreens Can Reveal User Identity: Capacitive Plethysmogram-Based Biometrics

Biometrics are widely used for user identification/authentication, but the fact has rarely been noticed that general capacitive touchscreens can reveal user identities by touch signals. This paper proposes a new biometric method with inherent liveness detection for reliable user recognition based on the cardiac signal captured by the capacitive touchscreen, namely Capacitive Plethysmogram (CPG). And a systematic framework is designed for CPG collection, processing, and exploitation to identify users. Specifically, since the finger usually forms capacitors with multiple sensing electrodes during touching, we can extract several CPG signals simultaneously from the screen output. Then we propose a series of preprocessing algorithms to filter CPG for signal quality enhancement. Finally, to further leverage filtered CPG signals and extract efficient features for identifying users, we build an encoder based on 3D attention CNN and metric learning. Experimental results demonstrate that the proposed method can achieve an average accuracy of 96.73%, FAR of 3.03%, and FRR of 7.35% in the laboratory environment, which reveals the potential of CPG for user privacy protection and data security on various devices laced with capacitive touchscreens.

Óvodás gyerekek névmodellje

Onomastic research typically focuses on the place-name recognition of adult name users. Thus, little research has focused on the place name vocabulary and naming competences of children. This paper presents the results of a survey using a complex methodology to characterise children’s place-name knowledge in a comprehensive way. Comprehensive interviews were conducted with ten children aged 6–7 years, as part of which children’s place-name knowledge, place-name competence and mental maps were assessed. The results show that children’s place-name knowledge is much narrower and spatially more limited than adults’, essentially covering only the inner area of their own municipality. They also know the macrotoponyms of other places, but the conceptual representations of these names are rather schematic. Children’s travel descriptions and mental maps are based on an egocentric perspective. The results show that the methodology can be applied successfully and provides a comprehensive picture of children’s spatial knowledge and place perception.

Beyond Objectivity in Audio Description

For many years, objectivity was considered a key norm of quality in audio description (AD), and it still features prominently in many AD guidelines. The primary motivation behind the concept is a recognition of users' autonomy in forming a mental image of the audiovisual content independently. Research has offered many alternative views on the issue that conceptualise the idea of objectivity/subjectivity as a continuum, emphasizing that different approaches can be appropriate depending on the given context of situation, giving rise to different AD styles that cater for a range of aesthetic preferences. Against this background, various cultural domains are experimenting with alternative approaches, challenging established AD practices and the assumptions it is built on. The articles included in this special issue all explore the various ways in which AD user needs can be met and illustrate the dynamic views that exist on objectivity, addressing key questions related to the operationalisation of objectivity, the challenge of interpretation for AD, the exploration of emerging subjective styles, and the concept of aesthetics of or in access.

Generation of keyboard handwriting during user authentication on mobile devices

Objectives. This article discusses a new way of generating keyboard handwriting using a touch keyboard for authentication in currently existing mobile systems.Methods. Due to the insufficient reliability of single password authentication, the proposal is to use it in combination with characteristics which correspond to handwriting on mobile devices. This article demonstrates the possibility of using individual user characteristics in the formulation of keyboard handwriting on devices with touch keyboards. The type of keyboard used affects the characteristics of keyboard handwriting, so this aspect can be used to improve password authentication reliability. The authentication process in the information environment can be supplemented with data on the nature of the impact on a touch keyboard. The use of the built-in 3D Touch function is also of interest. This is available when working on mobile devices and appliances equipped with a touch keyboard. The paper demonstrates that the use of one parameter only is insufficient for accurate authentication. The study proposes a method of determining an acceptable error range for both the touch force and the intermediate interval during authentication. For this purpose, the Laplace function which formulates the interval of each characteristic depending on the required probability of user recognition is used.Results. Touch force and the intermediate interval are sufficient to obtain the necessary characteristics, in order to formulate a refined user portrait depending on the user’s keyboard handwriting. Experimental statistics are given separately for an average sample of three different users depending on touch force. They also provide the results of authentication when using both standard deviations of pressing and the intervals when using the touch keyboard for the iOSXcode platform.Conclusions. The conclusion relates to the possibility of user authentication by keyboard handwriting, formulated on the basis of both the touch force on the keyboard symbols and intervals between pressing. Using the values of the sample mean and standard deviations allows authentication according to the required recognition probability.

DEVELOPMENT OF AN ACCESS CONTROL AND MANAGEMENT SYSTEM USING A SOFTWARE APPLICATION

The administration of educational institutions has several multi-storey buildings, in which it is necessary to carry out a set of technical and organizational measures to control visitors by security personnel. The lack of the possibility of registering students does not allow for the identification of their identity, which makes it difficult to investigate incidents of discipline violations and location. Registration of passport data of visitors with a small number of supervisory personnel is not suitable for the business processes of educational institutions. Measures to ensure the safety of visitors are carried out in educational institutions by selective control of visitors by security personnel, installation of remote control systems, video surveillance equipment. Functional modeling methods and programming were used to implement the identification and authentication method in the access system. The use of mobile devices, cards with a chip with a set of identification data implements the technology of their transmission, automating the access regime in educational institutions. The materials are compiled based on the results of research and development of an access control and management system using a MySQL, Python based software application. The article presents an analysis of the technology for developing a control and access system developed on the basis of Proximity. An experimental test of the software application obtained during the study was carried out, which implemented technologies for organizing communication with network devices, automatic detection of mobile devices in the database, user recognition by identification code. The software application interacts with other applications and chips, exchanges information, unidirectional communication with data stored on the server. A software application using PHP and MySQL database for student access control based on mobile devices can be used to determine the location of students who have passed the control.

A novel hybrid CNN-LSTM approach for assessing StackOverflow post quality

Abstract Maintaining the content quality on social media Q&A platforms is pivotal for user attraction and retention. Automating post quality assessment offers benefits such as reduced moderator workload, amplified community impact, enhanced expert user recognition, and importance to expert feedback. While existing approaches for post quality mainly employ binary classification, they often lack optimal feature selection. Our research introduces an automated system that categorizes features into textual, readability, format, and community dimensions. This system integrates 20 features belonging to the aforementioned categories, with a hybrid convolutional neural network–long short-term memory deep learning model for multi-class classification. Evaluation against baseline models and state-of-the-art methods demonstrates our system’s superiority, achieving a remarkable 21–23% accuracy enhancement. Furthermore, our system produced better results in terms of other metrics such as precision, recall, and F1 score.

Challenges of Health Informatics Systems in Primary Health Care

Health informatics systems are indispensable implements for cultivating the eminence and effectiveness of primary health care. Health informatics systems are formed of databases and figures regarding all clinical facilities. The utilization of health informatics systems helps both healthcare providers and patients in the management of disorders and disease risks and in developing well-being. However, there are a number of challenges to the application and practice of health informatics systems in the settings of primary health care. These challenges embrace; the absence of capital and money, the absence of technical skill, the absence of user recognition and approval, information quality problems, safety and confidentiality concerns, incorporation with other systems, and interoperability deficiency. Human, cultural, and financial barriers with deficiency of technical proficiency, deficiency of a relation between clinicians and patients, worries about information privacy, absence of administration guidelines, bureaucracy, lack of training, English language barrier, the degree of workload due to the time shortness during clinical settings, infrastructure concerns e.g. availability of the electricity, computers, and internet, also high turnover in IT among healthcare are also some challenges that should be addressed. These challenges can hold back the capacity of health informatics systems to comprehend their complete perspective in primary health care. To address these challenges, it is essential to provide sufficient capital and funds, offer practice and maintenance for users, overcome information quality concerns, apply safety and confidentiality procedures, incorporate health informatics systems with other systems, and endorse interoperability. By overcoming these challenges, health informatics systems can play a major role in developing the quality and efficiency of primary health care.

Research and reflection on college physical education classroom teaching based on SSGAN model

Abstract Teaching behavior recognition has a wide range of applications in the smart classroom and is one of the important means to achieve educational intelligence. To improve the performance of indoor teaching behavior recognition using CSI in complex scenes, this paper proposes an indoor teaching behavior recognition algorithm based on multi-feature fusion MLSTM by eliminating background noise to circumvent the influence of the experimental environment on CSI. To address the problem, the model cannot generalize in recognizing new users, and the labeled samples of new users are difficult to obtain in large quantities in a short period of time. In this paper, a new user recognition algorithm based on the SSGAN model is constructed, and then the input and output of MLSTM are modified as the discriminator of SSGAN to improve the recognition performance of the model for new users by semi-supervised learning. The recognition accuracy of the M-LSTM model on the sports, daily, and dance datasets is 0.985, 0.966, and 0.944, respectively, and the recognition accuracy of the SSGAN model on the three datasets is also around 90%, as verified by different experiments. The p-value is less than 0.05, and the student’s interest in physical education in the experimental group is 2.6 times higher than that in the control group. Therefore, the model proposed in this paper has good practicality.

Studying the Visual Representation of Microgestures

The representations of microgestures are essentials for researchers presenting their results through academic papers and system designers proposing tutorials to novice users. However, those representations remain disparate and inconsistent. As a first attempt to investigate how to best graphically represent microgestures, we created 21 designs, each depicting static and dynamic versions of 4 commonly used microgestures (tap, swipe, flex and hold). We first studied these designs in a quantitative online experiment with 45 participants. We then conducted a qualitative laboratory experiment in Augmented Reality with 16 participants. Based on the results, we provide design guidelines on which elements of a microgesture should be represented and how. In particular, it is recommended to represent the actuator and the trajectory of a microgesture. Also, although preferred by users, dynamic representations are not considered better than their static counterparts for depicting a microgesture and do not necessarily result in a better user recognition.

AdaML: An Adaptive Meta-Learning model based on user relevance for user cold-start recommendation

User cold-start recommendation is a critical problem due to the sparse user–item​ interactions of new users. To address this issue, an emerging popular direction employs the meta-learning paradigm, which can quickly adapt to new tasks with sparsely labeled data. Early meta-learning recommendation works use the same global prior to initialize the parameters for each user’s preference learning task, which may be insufficient to discern users’ complex behaviors and leads to poor generalization. As a result, recent works have proposed the concept of relevance-based meta-learning recommendation, which ensures that similar priors are locally shared by related users with similar preferences, and have yielded promising results. However, as an important factor influencing the performance of relevance-based meta-learning recommendation, related user information is still not sufficiently explored and exploited in existing works. To that end, in this paper, we present a novel adaptive meta-learning model based on user relevance (AdaML) to further improve the user cold-start recommendations. Specifically, AdaML firstly proposes a related user recognition component, which recognizes related users with similar preferences by integrating the meta-path-based and the clustering-based strategies. The retrieved high-quality information of related users is then effectively utilized by a well-designed user relevance enhanced recommender, on one hand, to enhance user representations and tailor relevance-specific priors from the globally shared prior, on the other. Finally, extensive experiments on several public datasets demonstrate AdaML’s superior performance over state-of-the-art approaches in the user cold-start scenario.

An Intelligent System for Fingerprint Recognition and Verification Using Dilated Convolutional Neural Network and Extreme Learning Machine

Several applications currently utilize fingerprint-based user recognition and authentication, but obtaining complete accuracy (eliminating false matches) is still challenging. Prior to feature extraction, improper image alignment was one of the causes of this problem. This paper proposes a novel fingerprint recognition and verification system using a dilated convolutional neural network (DICNN) and a weighted and bias-optimized extreme learning machine (WBELM). Image preprocessing, feature extraction, and fingerprint recognition are the main phases of the proposed work. To begin, image preprocessing is performed using bilateral filtering to suppress the noise in the gathered images. After that, the proposed system uses DICNN to extract features from the preprocessed images. From the extracted features, fingerprint recognition is done using WBELM. This study conducted experiments on the publicly available FVC 2004 dataset. The proposed model achieves 98.54 percent recognition accuracy with an elapsed time of 584 ms.

Achieving Information Security by multi-Modal Iris-Retina Biometric Approach Using Improved Mask R-CNN

The need for reliable user recognition (identification/authentication) techniques has grown in response to heightened security concerns and accelerated advances in networking, communication, and mobility. Biometrics, defined as the science of recognizing an individual based on his or her physical or behavioral characteristics, is gaining recognition as a method for determining an individual's identity. Various commercial, civilian, and forensic applications now use biometric systems to establish identity. The purpose of this paper is to design an efficient multimodal biometric system based on iris and retinal features to assure accurate human recognition and improve the accuracy of recognition using deep learning techniques. Deep learning models were tested using retinographies and iris images acquired from the MESSIDOR and CASIA-IrisV1 databases for the same person. The Iris region was segmented from the image using the custom Mask R-CNN method, and the unique blood vessels were segmented from retinal images of the same person using principal curvature. Then, in order to aid precise recognition, they optimally extract significant information from the segmented images of the iris and retina. The suggested model attained 98% accuracy, 98.1% recall, and 98.1% precision. It has been discovered that using a custom Mask R-CNN approach on Iris-Retina images improves efficiency and accuracy in person recognition.

EdBB-Demo: Biometrics and Behavior Analysis for Online Educational Platforms

We present edBB-Demo, a demonstrator of an AI-powered research platform for student monitoring in remote education. The edBB platform aims to study the challenges associated to user recognition and behavior understanding in digital platforms. This platform has been developed for data collection, acquiring signals from a variety of sensors including keyboard, mouse, webcam, microphone, smartwatch, and an Electroencephalography band. The information captured from the sensors during the student sessions is modelled in a multimodal learning framework. The demonstrator includes: i) Biometric user authentication in an unsupervised environment; ii) Human action recognition based on remote video analysis; iii) Heart rate estimation from webcam video; and iv) Attention level estimation from facial expression analysis.

Emotion recognition of social media users based on deep learning.

Issues with sentiment analysis in social media include neglecting the long-distance semantic link of emotional features, failing to capture the feature words with emotional hue effectively, and depending excessively on manual annotation. This research provides a user emotion recognition model to achieve the emotional analysis of microblog public opinion events. Three types of inspiring text, "joy," "anger," and "sadness," are obtained by the data collecting and data preprocessing of micro-blog public opinion event comment text. Then, an algorithm using the linear discriminant analysis (LDA) model, emotion dictionary, and manual annotation is created to extract emotional feature words. The captured motivational text is converted into a word vector using Word2vec. After gathering the long-distance semantic data with bidirectional long short-term memories (BiLSTM) and convolutional neural networks (CNN) extract the text's key characteristics to finish the emotion categorization. The test results demonstrate an average increase in F1 value of 3.66 percent for six machine learning models and an average increase in F1 value of 1.84 percent for seven deep learning models. The suggested model performs better at identifying the emotions of social media users than the current machine learning and deep learning methods.

Detecting unregistered users through semi-supervised anomaly detection with similarity datasets

Recent research has focused on exploring systems that incorporate anomaly detection models to automate the addition of users in user recognition systems. Anomaly detection, a method used to distinguish between new and existing users by identifying abnormal images, has gained significant attention. Researchers have been actively investigating the Semi-Supervised Learning method, which utilizes only existing user data to differentiate between existing and new users. However, existing semi-supervised learning based anomaly detection models exhibit high performance on datasets with low similarity but experience a sharp decline in performance on datasets with high similarity. Furthermore, their large model size makes it challenging to execute them on edge nodes. To address these limitations, this paper proposes a model that can be executed on edge nodes and guarantees good performance on both low and high similarity datasets. The proposed model utilizes the LeNet-5, a user recognition model with fewer weights and multiple images as input, for classifying new users. This study compared the existing anomaly detection models with the proposed model using three datasets with varying similarities. The performance evaluation involved comparing the accuracy, ROC curve, and AUC of each model on a training server. Subsequently, the top three models were optimized for execution on the edge node (STM32F207ZG MCU) and further evaluated by comparing their accuracy, inference speed, and model size. The results revealed that the proposed model achieved an approximate 53% improvement in accuracy compared to the existing anomaly detection models. Furthermore, when executed on the edge node, the proposed model demonstrated significant memory savings, with a maximum reduction of approximately 530% and approximately 40% reduction in flash memory usage compared to the existing models.

Reliable Multi-Factor User Authentication With One Single Finger Swipe

Multi-factor user authentication becomes increasingly popular due to its superior security comparing with single-factor user authentication. However, existing multi-factor user authentication methods usually require multiple interactions between users and different authentication components when inputting the multiple factors, leading to extra overhead and bad user experience. In this paper, we propose a secure and user-friendly multi-factor user authentication system named <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BioDraw</i> . It utilizes four categories of biometrics (impedance, geometry, behavior, and composition) of human hand plus the pattern-based password to identify and authenticate users. User only needs to draw a pattern on a radio frequency identification tag array, while four biometrics can be collected simultaneously. Specifically, we first design a gradient-based pattern recognition algorithm to precisely extract user’s secret pattern. Then, a convolutional neural network-and long short-term memory-based classifier is utilized for user recognition. Furthermore, to guarantee the systemic security, an anti-replay method called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Binary ALOHA</i> is proposed to detect replayed signals. We conduct extensive experiments with 30 volunteers. The experiment results show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BioDraw</i> can achieve high authentication accuracy (with a 2% <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$-$</tex-math> </inline-formula> false reject rate) and is effective in defending against various attacks.

An Experimental Analysis on Multicepstral Projection Representation Strategies for Dysphonia Detection.

Biometrics-based authentication has become the most well-established form of user recognition in systems that demand a certain level of security. For example, the most commonplace social activities stand out, such as access to the work environment or to one's own bank account. Among all biometrics, voice receives special attention due to factors such as ease of collection, the low cost of reading devices, and the high quantity of literature and software packages available for use. However, these biometrics may have the ability to represent the individual impaired by the phenomenon known as dysphonia, which consists of a change in the sound signal due to some disease that acts on the vocal apparatus. As a consequence, for example, a user with the flu may not be properly authenticated by the recognition system. Therefore, it is important that automatic voice dysphonia detection techniques be developed. In this work, we propose a new framework based on the representation of the voice signal by the multiple projection of cepstral coefficients to promote the detection of dysphonic alterations in the voice through machine learning techniques. Most of the best-known cepstral coefficient extraction techniques in the literature are mapped and analyzed separately and together with measures related to the fundamental frequency of the voice signal, and its representation capacity is evaluated on three classifiers. Finally, the experiments on a subset of the Saarbruecken Voice Database prove the effectiveness of the proposed material in detecting the presence of dysphonia in the voice.

Emotion Recognition using Artificial Intelligence

This Paper is intended to explore research done on identifying research done on the identifying the face recognition of users and play music according to the mood of the users by using AI and ML

RAIF: A deep learning‐based architecture for multi‐modal aesthetic biometric system

AbstractHuman aesthetics play a significant role in video game development, emotional‐aware robot design, online recommender systems, digital human, and other domains of research focusing on human‐computer interactions. Social network user recognition based on aesthetic preferences is an emerging research domain. In this paper, a novel deep learning architecture is proposed for multi‐modal audio‐visual person identification that combines audio and visual aesthetic features. A pre‐trained ResNet architecture is utilized to extract high‐level features from a set of user‐preferred audio and image samples. A novel deep learning‐based fusion technique called residual‐aided intermediate fusion (RAIF) is introduced in order to effectively merge the audio and visual features. The proposed RAIF method achieved an accuracy of 98% and a loss of 0.01 on a proprietary multi‐modal dataset, indicating its effectiveness in fusing audio and visual information.