Hand gesture recognition has become an indispensable part of human-computer interaction. It supports intuitive, contactless, and accessible means of communication. This work proposes a robust deep learning-based system for the real-time static recognition of hand gestures to assist communication with the deaf and hard-of-hearing. We propose hybrid architecture that consists of the Swin Transformer for contextual attention, ResNet34 for spatial feature extraction, and a BiLSTM layer for temporal understanding. The model was trained and evaluated using a subset of the HaGRID dataset, consisting of 18 different classes of hand gestures. It achieves a training accuracy rate of 99.3% and a testing accuracy rate of 98.03%, thereby demonstrating its excellence in performance amid different lighting and backgrounds. Our approach proposes a scalable solution that can be integrated into assistive technology for gesture-to-text or gesture-to-speech conversion, thus bridging the communication gap for deaf individuals.
In this paper, we have estimated the reliability of a system composed of elements connected in series and subject to the same stress, by the nonparametric kernel estimation method using lognormal kernel. This kernel is chosen for its advantageous properties, including an optimal convergence rate for the mean integrated squared error, non-negativity, free from boundary bias, and naturally varying in shape. Asymptotic properties such as bias, variance, and mean squared error have been established for the proposed estimator. Furthermore, we address the selection of the optimal bandwidth parameter an essential aspect of kernel estimation using both the rule of thumb and the unbiased cross validation techniques. Finally, a simulation study is carried out to highlight the performance of the system reliability estimator, based on lognormal kernel and to compare the two bandwidth selection techniques to determine the best one.
This paper investigates a Bivariate mixture of the student’s t and Normal distributions, referred to as the t-Normal distribution, where the marginals are univariate Student’s t and Normal distributions, respectively. The study derives the conditional distributions along with their associated constants. It also presents several generating functions, including the moment, cumulant, characteristic, hazard, survival, cumulative hazard functions, and Shannon’s differential entropy. Three-dimensional probability surfaces illustrate the shape of the t-Normal density. Special cases of this distribution include mixtures involving logarithmic, logit, and hyperbolic sine transformations. Furthermore, the paper derives the maximum likelihood estimators for the parameters, calculates the elements of the Fisher Information matrix and explores parameter estimation through a nonlinear programming approach.
This project develops an AI-powered Smart Traffic Light System utilizing YOLOv8 (You Only Look Once) for real-time vehicle detection to improve and better traffic management. The system dynamically adjusts green light durations based on the number of vehicles detected. If there are five or less then five vehicle detected, the green light duration is calculated as the number of vehicles multiplied by five seconds. If there are more than five vehicles, a fixed 40-second green light duration is applied. Additionally, the system prioritizes emergency vehicles, immediately providing them with a 40-second green light when detected, ensuring faster emergency response times. The system was implemented using Python, OpenCV, and Ultralytics YOLOv8, with inference accelerated through GPU processing. The results demonstrate the effectiveness of the system in real-world traffic management, with improvements in both traffic flow and emergency vehicle prioritization.
The metaverse can extend the limits of reality by using augmented and virtual reality technology. It allows individuals to connect seamlessly in genuine and artificial environments by employing avatars and holographic pictures. Simulations and virtual reality games are considered precursors of the metaverse, offering insight into the potential social and economic implications of a fully realized and universally accessible metaverse. Differentiating between the exaggerated promotion of the “meta” redesign and the actual realities of the moment can be difficult due to the influential promotion by major technology companies. They present the metaverse as a significant catalyst that will benefit people’s professional pursuits, leisure activities, and social connections. The emerging concept of a gradual merging of digital and physical realms could potentially have a transformative impact on various aspects of human activities, including work performance, communication with organizations and individuals, and the formation of collective memories. Although the necessary technology resources and infrastructure for creating new real-world worlds that human avatars can explore through gadgets are not yet available on a large scale, scholars are increasingly interested in the transformative potential of the metaverse. Normal social interaction parameters have farreaching effects on many fields, including commerce, healthcare, academia, and society at large. Confidence, secrecy, prejudice, deceptive data, legal compliance, and mental health difficulties related to reliance and the impact on vulnerable populations are all factors to consider. This research delves into several subjects related to the metaverse and its transformative impact by using an impartial narrative and a comprehensive approach that draws on the expertise of specialists from different disciplinary backgrounds. The report concludes by proposing a comprehensive plan for further investigation that will be beneficial to scholars, practitioners, and decision-makers on a global scale.
The rapid growth of regional language news videos faces major challenges for efficient indexing and retrieval, particularly for regional languages such as Gujarati, which is spoken by 55 to 62 million people worldwide. Processing every frame in videos is time-consuming and redundant; therefore, accurate shot boundary detection followed by effective key-frame selection is important for text-based video retrieval. This research presents an adaptive video segmentation framework for Gujarati news videos that integrates shot boundary detection with key frames selected using statistical methods. Initially, consecutive video frames are processed category-wise, and six frame-level features are extracted from the videos, such as PDM, CDM, HBA, Gabor response, ECR, and EED. These features are then normalized and fused using a weighted approach. Mean and standard deviation based adaptive thresholds are then applied to detect both abrupt and gradual shot transitions. For the detected shots, a key frame has been extracted from shots containing Gujarati text based on the Ground truth table. This approach uses entropy and edge density to capture detailed information and structural variation. Adaptive thresholds derived from their mean values are used to preserve descriptive frames with high textual significance. Experimental evaluation on TV9 Gujarati news video datasets shows that the proposed framework effectively reduces redundant frames while preserving semantically significant content key frames, which are suitable for efficient Gujarati text-based video retrieval. Total no. of shots reduced is 2158 from 80675, and total key frames are 901, with the lowest accuracy in the weather category, with 81.0714 % and the highest recorded as 88.00 % for the Cricket category.
The Indian organic food market is growing at a very high rate owing to increase in environmental concerns, health concerns as well as due to the shift towards sustainable consumption. This study explores the most crucial psychological and value driven motivation that influences the intention of the consumers to purchase the organic food. The study is appropriate to posit and empirically verify that the model comprises of the perceived quality, the implications and personal worth and green trust as an aspect of consumer attitude which conversely affects consumer buying intention. Both the Theory of planned behaviour and end chain theory are the foundation of this model. Data has been gathered using a structured questionnaire built around 5-point Likert and entailed acquiring the responses of 190 people who can be described as organic food customers, and it was conducted using convenience and snowball sampling. The data analyses were confirmatory factor analyses using structural equation modeling. In the results, we see that all the variables have an impact on the customer attitude in which the effects seem to be the best predictor, then perceived qualities then personal values. Purchase intention is then highly influenced by attitude. These results highlighted the role of trust and values as a critical factor in determining the purchase decision of the consumers to consume organic food. This research can provide them with hints to the marketers and to the policymakers to develop the campaign founded on the principles of trust and value-oriented to motivate people to consume organic food.
The global transition toward sustainable energy systems is accelerating due to increasing concerns about climate change, carbon emissions and energy security. Renewable energy sources such as solar and wind provide environmentally sustainable alternatives to fossil fuels; however, their inherent intermittency, stochastic behavior and environmental dependency introduce significant operational and planning challenges for modern energy systems. Artificial intelligence (AI) and deep learning techniques have recently emerged as effective solutions for addressing these challenges through accurate forecasting, intelligent control and optimized energy management. This review paper presents a systematic and comprehensive analysis of recent advancements in AI-driven approaches for renewable energy assessment and management. A structured literature review methodology was adopted to analyze 30 peer-reviewed research articles obtained from major scientific databases, including IEEE Xplore, ScienceDirect and Google Scholar. The selected studies were categorized into four major application domains: solar energy forecasting, wind energy optimization, hydropower resource management and hybrid renewable energy systems. The review critically evaluates widely adopted machine learning and deep learning models such as Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) networks, and hybrid AI frameworks, highlighting their predictive capabilities and operational benefits. Furthermore, this study identifies key research gaps related to data availability, model interpretability and scalability while outlining future research directions for developing intelligent, reliable, and sustainable renewable energy systems.
The steel sector is a fundamental industry in the global economy, facing challenges in both safety aspects and the need to increase competitiveness in response to growing environmental regulations and global market pressures. This article, based on information provided by various sources, reflects on the multifaceted role of digital technology (including technology related to Industry 4.0) in addressing these challenges. Digital tools result in safer environments. Digital technologies help make workplaces safer by enabling automated monitoring systems, such as Computer Vision, for detecting potential hazards and the use of robotics to keep workers away from dangerous situations. Moreover, solutions like AI-powered Predictive Maintenance enhance the reliability of your equipment, which is extremely important for security in high-risk environments and also to enhance operational efficiency. At the same time, digital innovations are important drivers of competitiveness by enabling more energy and resource-efficient production, incentivising decarbonisation, and improving supply chain management. Despite the promise, the implementation of such solutions will be contingent on research knowledge gaps regarding the use and application of solutions in practice, as well as the training of the workforce.
The financial Internet is making our country’s digital divide, especially with no infrastructural developments. Digitalization could connect to deliver worldwide and combined access to high-quality financial services. Customers are looking for less fraudulent activity to make transfers less costly, faster, instantaneous, and secure and make the system more efficient. The digital divide touches every area of life, including literacy, wellness, security, access to financial services, etc. Therefore, a developing country, India, must focus on inclusion and equitable growth. The researcher could identify through the research where the weightage is to be given, that is, timely intervention in technological access and proficiency within the sample population. The study found no significant association between education levels and digital literacy on the digital divide in the country. The expansion of financial inclusion has also been severely hindered by a lack of financial knowledge, resulting in a lower acceptance rate. The best way to close the digital gap in financial inclusion are limited to appropriate practices and education.
The present study investigates the Efficiency and Financial performance of Indian general insurers during the period 2016-2022. This study measures the level of efficiency of these firms by applying Data Envelopment Analysis (DEA) to identify trends in operating effectiveness, profitability, and solvency. Findings indicate that while public insurers flounder due to higher claims ratios and operating costs, private insurers generally perform well with respect to solvency and investment productivity. The practical advice on financial sustainability and regulatory compliance can be improved in the insurance industry of India. Benefit of stakeholders by assisting them to enhance the performance of the sector.
As industries evolve toward digital transformation, the integration of traditional quality management tools with emerging technologies becomes imperative. This paper presents a novel framework that synergizes Six Sigma methodologies with Artificial Intelligence (AI) to align with the principles of Quality 4.0. By leveraging machine learning, real-time analytics, and big data, the proposed model enhances the DMAIC (Define, Measure, Analyze, Improve, Control) cycle for continuous improvement in smart manufacturing and service environments. The framework introduces AI-powered tools for root cause analysis, predictive quality, and intelligent decision-making, thereby reducing process variation and enhancing operational efficiency. Case studies and simulations demonstrate the effectiveness of this integrated approach in driving superior quality outcomes and enabling agile responses to dynamic market demands. This research bridges the gap between statistical process control and AI-driven quality assurance, offering a scalable pathway for organizations to achieve excellence in the Industry 4.0 era.
In this study, we compare the performance of various machine learning algorithms through simulations conducted on datasets with different distributional characteristics. Widely used models such as Logistic Regression, Naive Bayes, k-Nearest Neighbors, Decision Trees, Support Vector Machines, Random Forest, AdaBoost, Gradient Boosting, and XGBoost are examined. The results indicate that the performance of classification algorithms is strongly influenced by the underlying data distribution and class balance. Specifically, Logistic Regression, Naive Bayes, and Support Vector Machines achieved high accuracy on normally distributed datasets, whereas tree-based methods such as Random Forest, Gradient Boosting, and XGBoost demonstrated greater robustness and superior outcomes on non-normal distributions and imbalanced class structures.
The paper derives the fiducial limits for a determinant D of order 4, filled with independently and identically distributed (i.i.d) Exponential variates, using Chebyshev’s inequality and then, comparing the derived result with the fiducial limits for D of order 2 and 3 obtained in previous research, generalizes the result for order n. If D is of order n with i.i.d. Exponential variates (θ), P( −k √((n+1)! / θ^(2n)) < D < k √((n+1)! / θ^(2n)) ) ≥ (1 − 1/k²), where k is a positive integer with E(D) = 0 and Var(D) = (n+1)! / θ^(2n).Several applications are pointed out. The study may be extended to other probability distributions.
Indeed, in e-commerce as an evolving industry, applying generative AI has been regarded as the key to the revolution of the entire pricing strategy. With the increase in the amount of available data and computational resources, the application of dynamic pricing strategy is more characteristic for e-commerce organizations. This article will concentrate on the comparison of the generative AI-based dynamic pricing and the traditional price techniques for instances specifically based on their effects on the efficiencies of revenue management and customer experience. The appearance of the term dynamic pricing in the sphere of e-commerce was possible due to the fact that large amount of data could be analyzed to find such sources of steady and predictable revenue and demand forecast was being made. However, in earlier research on these models, the most used variable was the Price Elasticity of Demand constructed from historical information; however, emerging complexities of consumers and markets required more studies on pricing models incorporating generative AI. This theoretical as well as practical research paper aims to underlie extended knowledge about the generative AI dynamic pricing strategies and assess their efficiency in comparison with the traditional price strategies. Such elements of costbenefit analysis as cost efficiencies, customer satisfaction levels, opportunities for increasing revenues etc for changing conditions are pinpointed. This research uses literature review and the analysis of cases to examine the key concepts and practical implementation of the generative AI-driven dynamic pricing in e-commerce. It analyses a critical set of factors specifically, personalization strategies, sensitivity to price changes, and perceived value regarding consumer behavior whilst making a purchase in dynamic price context. The study also compares the effectiveness of dynamic pricing systems based on generative AI with traditional models, including fixed, time-based, and competition-based pricing schemes. Further, in this work, it has been illustrated that e-commerce companies can gain significant benefits by integrating generative AI into dynamic pricing frameworks, such as higher revenue, enhanced customer experience.
The widespread adoption of digital HR has created a bandwidth for advanced tools like HR analytics incorporating data-led approaches to improve employee engagement. This paper examines the role of HR analytics on software employee engagement in Hyderabad, Telangana which is part of India to one of the IT hubs. Hence, the research combines quantitative analysis through survey data with qualitative insights from interviews to assess how HR analytics impacts employee engagement (motivation, satisfaction, and retention). A structured survey was circulated among 300 mid- to large-scale IT company employees in Hyderabad. We developed a questionnaire that examined employees’ perceptions of HR analytics practices to govern how HR analytics are aligned closely with organizational needs, and whether or not these affect key engagement metrics. Regression analysis and other statistical tools were used to analyse the quantitative data. In addition, semi-structured interviews with HR managers and software employees explored case-shifting efforts used with HR analytics in engagement initiatives because of many encouraging effects from the literature around this area. The findings indicate that a higher level of effective application of HR analytics leads to increased employee engagement. HR analytics tools helped bring personalized experiences to employees, shriveled workforce sentiment, and predicted engagement risks. More transparent use of people analytics also fosters trust among employees [14]. Predictive analytics have long played a role in helping catch the early signs of disengagement, and act before losing an employee, this study reaffirms that potential. A mix of customized training programs, growth path maps, and data-oriented recognition systems were highlighted as top engagement drivers. On the positive side, the study identified opportunities such as data privacy issues and insufficient technical expertise among HR professionals. If appropriately managed, such anxiety can lead to the negation of benefits that could have been generated from HR analytics for the organization since a vast majority of employees expressed hesitance over how their personal data was being sourced and used. In addition to this, the organizations also had to go through a change in mentality where HR analytics role (i.e. using data for people’s decision-making) was formalized which required training of and trust in HR personnel on how to interpret and act based on analytic outputs. This study emphasizes the importance of HR analytics in creating an engaged workforce in the software sector. Offering effective insights and customizing HR practices to match employee requirements goes a long way in improving engagement numbers as well. The results support the implementation of strong HR analytics systems as well as high transparency regarding potential privacy breaches and their management to establish trust from the employee perspective. Longitudinal and field research with HR analytics: Future studies could examine the lagging effects of HR metrics on employee performance (loyalty) as well as on profitability.
This research paper focuses on examining the behaviour of stock markets and investigating whether global stocks exhibit the property of long memory. Several methodologies, such as aggregated variance, rescaled-range analysis, and the periodogram, are employed to test this hypothesis. The criteria for determining the presence of long memory involve the point estimation of the Hurst exponent. Beyond point estimation, the return series and volatility series are further analysed using rolling samples to observe the dynamic nature of long memory over the studied period. Our findings suggest that market behaviour evolves over time, demonstrating that a single static Hurst exponent is insufficient to assess the overall efficiency of the market.
Telematics, the integration of telecommunications and informatics, has emerged as a disruptive and innovative force in the automotive insurance sector. This paper presents a bibliometric analysis of the application of telematics within this domain. The analysis focuses on the co-occurrence and connections between telematics and automotive insurance, offering insights into trends, advancements, and future directions. By mapping the connections in the literature, the paper aims to provide conceptual clarity and further understanding of how telematics is shaping the automotive insurance landscape. The findings highlight the role of telematics in transforming traditional risk assessment models and customizing insurance premiums based on real-time driver behaviour.
Radial basis function (RBF) networks, deep neural networks (DNN) with Adam optimization, spline interpolation, polynomial approximation, and DNN with Levenberg- Marquardt (LM) optimization are five sophisticated techniques used in this work to build a novel universal linearization framework. Through adaptive mode selection of the most appropriate technique depending on the unique characteristics of the sensor data, the proposed system achieves higher accuracy and robustness in handling diverse nonlinearities. Experimental results demonstrate remarkable improvement in linearization performance for various kinds of thermocouple sensors, witnessing the usability and efficiency of the framework for real-time applications.
The present study examines the role of capital structure on financial performance of energy sector companies in a major emerging economy. By concentrating on the moderating effect of corporate governance practices on ownership structure and corporate performance, it closes a gap in the research. It uses energy companies financial data for a period of 10 years (2014-2023), employing panel regression approach using fixed-effect estimation. Debt-equity ratio, debt-asset ratio has been considered for measuring capital structure whereas return on assets has been taken as proxy of financial performance. The results were validated using GMM model (Generalized Method of Moments) to control for potential endogeneity. Corporate governance has been measured by size of Board, , CEO duality, board independence, and size of audit committee. Based on regression results, the study finds a substantial detrimental impact of capital structure on performance of business. Further, the moderation analysis has revealed mixed results as board size and independence positively moderates performance of firm and capital structure whereas CEO duality and audit committee have negative moderating effect. The current study provides significant implications for management and extends literature on debt financing and corporate performance particularly the underexplored corporate governance’s role.
