Abstract This study examines the reintegration barriers and prospects affecting the emotional and psychological well-being of Gulf returnees in Kerala. The research analyses the difficulties encountered by return migrants across five key dimensions: financial constraints, limited government support (GS), family reintegration challenges, societal reintegration barriers, and career transition difficulties. Primary data were collected from 768 return migrants across selected districts in Kerala identified through the Kerala Migration Survey (2018). Using Structural Equation Modeling, the study identifies the most significant factors affecting the reintegration process. The findings reveal that economic difficulties—particularly debt, declining income, and reduced savings—and societal reintegration challenges represent the strongest barriers. Societal barriers mainly include difficulty rebuilding social networks, perceived social discrimination, and limited participation in local community organizations. Government program access also emerged as a major concern due to lack of awareness and bureaucratic constraints. In contrast, family reintegration and career transition factors showed comparatively weaker influence on overall reintegration challenges. The demographic analysis indicates that most returnees are middle-aged Muslim men who worked in low- to medium-skilled occupations in Gulf countries for more than a decade. The study highlights how economic insecurity and social marginalization indirectly affect the emotional and psychological well-being of return migrants. These findings underscore the need for targeted policy interventions, including improved financial assistance, accessible GS programs, and community-based reintegration initiatives to facilitate sustainable socio-economic and psychological adjustment for Gulf returnees in Kerala.
Abstract The growing use of IoT devices in healthcare has increased medical data security challenges, as traditional access control models like role-based access control (RBAC), attribute-based access control (ABAC), and mandatory access control (MAC) lack flexibility and context awareness. The healthcare Internet of Things (IoT) faces challenges such as limited scalability, lack of dynamic access controls, and security risks from a centralized point of vulnerability. The research aims to develop a resilient, decentralized access control solution that provides secure, time-sensitive permissions in healthcare IoT systems. A hybrid RBAC-ABAC model, built on blockchain and integrated with interplanetary file system (IPFS), enables secure data sharing across healthcare IoT devices, outperforming current models in accuracy, latency, and scalability. A dual smart contract–IPFS mechanism enables adaptive access control. Results show 96.5% precision, 3.2 ms policy evaluation, 120 ms response time, 74 TPS, low cost (US$2.1), and 45–52 MB memory use. It outperforms ABAC, RBAC, and MAC, offering scalable, efficient security for healthcare IoT data sharing.
Abstract Neuromorphic computing is an emerging transformative approach inspired by the structure and functioning of biological neural networks toward enhancing the efficiency and adaptability of artificial intelligence systems. This work explores the use of spiking neural networks (SNNs), memristive devices and novel two-dimensional materials to advance neuromorphic computing technologies. We analyzed these systems using comparative experimental evaluation and discussed applications such as remote sensing scene classification and sensory processing based on recent neuromorphic computing studies. Results indicated reduced inference latency characteristics for SNN-based neuromorphic systems compared with conventional neural architectures under equivalent evaluation settings. The biomimetic neuromorphic sensory system based on electrolyte-gated transistors demonstrated approximately 15% lower estimated energy consumption compared with conventional processing approaches reported in related studies. Memristive neuromorphic devices demonstrated improved synaptic adaptation characteristics, indicating enhanced learning and plasticity behavior in comparison with conventional implementations reported in related studies. The study further discusses future prospects of neuromorphic computing using advanced materials and adaptive architectures.
Abstract The recent improvements of current RGB-D salient object detection models have achieved better results by leveraging the depth modality in a convolutional neural network. Most existing approaches use a multi-stream architecture with enhanced feature fusion to subsequently determine saliency. The inconsistency of salient regions may result from the limitations of the acquisition devices and the low-depth images. Multi-horizon feature characteristics and complementary modalities are essential for resolving these problems. Even in cluttered backgrounds, complementary and multi-horizon features across several horizons must be taken into consideration since they provide structural and contextual information about the salient object. To maximize important features in low-depth images, the proposed model addresses these constraints. The multi-horizon features are optimized from the RGB and depth streams by using the proposed backbone network with the feature aggregation and optimization (FAO) module. The FAO module, which optimizes feature extraction by utilizing complementary features from the depth stream, is used to enhance the RGB stream. The proposed model investigates multi-horizon features between multi-resolution and multi-stage complementary features using a multi-horizon reconstruction module. To forecast optimal saliency, the FAO module combines data at multiple levels. A guiding framework for maximizing salient regions and reducing non-salient ones throughout the fusion process is provided by attention maps at various phases. Current evaluation metrics from the six publicly accessible difficult RGB-D datasets are used in the experimental investigation. Furthermore, we compare our findings with those of 18 other cutting-edge methods that have demonstrated encouraging performance on the above challenges.
Abstract Neuromorphic computing is an emerging brain-inspired paradigm that enables efficient information processing by integrating computation and memory in massively parallel, event-driven architectures. Unlike traditional von Neumann systems, it is capable of alleviating—though not necessarily eliminating—constraints related to memory, power inefficiency and real-time learning latency, particularly in applications that are sparse, event-driven and time-sensitive designed for spike-based or in-memory computing; however, in applications that are dense, are throughput-oriented and involve batch processing, traditional computing systems may be more efficient. This review highlights key neuromorphic models, including spiking neural networks and biologically inspired learning rules such as spike-timing-dependent plasticity, which support adaptive and energy-efficient computation. It also discusses the evolution of neuromorphic hardware from digital platforms to analog and mixed-signal implementations using emerging technologies such as memristors and phase-change memory. Major applications and key challenges in neuromorphic computing are briefly outlined.
Abstract Melanoma, a more aggressive class of skin cancer, develops from melanocytes, pigment-producing cells in the skin. Late diagnosis increases the risk of cancer spreading to other parts of the body, thereby decreasing the patient’s survival rate. An early diagnosis of melanoma cancer not only increases the survival rate but also improves the disease prognosis. Reduced medical costs incurred during cancer treatment are one of the major advantages of early cancer diagnosis. This research study proposes and implements a novel framework that leverages the enhanced super resolution generative adversarial network (ESRGAN) for pre-processing. The segmentation of the region of interest in preprocessed high-resolution images is carried out by density-based clustering algorithm with the Salp optimization technique. In the segmentation phase, the Salp Optimization algorithm computes the optimum cluster center for enhancing the performance of the density-based clustering algorithm. The classifier adopted in this study is a custom-built convolutional neural network model. The proposed convolutional neural network model with set parameters demonstrates a high accuracy and recall rate, deferring the need to adopt time-consuming and complex deep learning models. The experimental study evaluates the performance of the proposed methodology and emphasizes that the implemented diagnosis tool shows accuracy of 94.5% in melanoma diagnosis. The peak signal noise ratio (PSNR) and structural similarity index metrics are evaluated for the ESRGAN-generated images, showing significant increase in image resolution with the method adopted.
Abstract The rapid expansion of Internet of Things (IoT) networks introduces critical security challenges, particularly during data transmission. This manuscript presents a novel lightweight cryptography framework that integrates the Firefly Optimization with Elliptic Curve Cryptography (FA-ECC) to optimize key management and enhance data security. Additionally, a dense recurrent neural network (DRNN) module predicts and blocks malicious data entries, ensuring only legitimate data reaches the cloud. The proposed framework simultaneously improves security, efficiency and quality of service in IoT networks. Experimental evaluations demonstrate that the FA-ECC and DRNN integration effectively reduces energy consumption, transmission delay and vulnerability to attacks, establishing a secure and optimized data exchange mechanism for IoT environments.
Abstract Sustainable agriculture in India is facing numerous challenges, including an increase in atmospheric temperature, resource depletion, and higher productivity to feed a growing population. The amalgamation of Artificial Intelligence (AI) and the Internet of Things (IoT) offers revolutionary solutions for addressing these challenges and promoting sustainable farming practices. This promise of AI and IoT technologies enhancing agricultural efficiency, reducing waste, and optimizing resource management in India is discussed. Indian farmers can utilize the strength of AI to know current information on soil health, weather patterns, irrigation needs, and pest control. The paper also discusses the barriers to adoption, including technological infrastructure, affordability, and the need for digital literacy among farmers. Furthermore, it highlights the integrated application of AI and IoT technologies, along with comprehensive results and a gap analysis, as a strategic approach to advancing sustainable agriculture and ensuring food security in India. The findings underscore the potential of these technologies to enhance environmental stewardship and strengthen economic resilience within the agricultural sector.
Abstract Cardiovascular disease (CVD) is one of the leading causes of mortality worldwide, necessitating the development of accurate and efficient predictive models for early diagnosis and intervention. However, the high dimensionality, redundancy, and heterogeneity of clinical data often limit the performance of conventional machine learning (ML) approaches. To address these challenges, this study proposes a hybrid feature selection with dimensionality reduction framework for CVD detection (HFSDR-CVD). The proposed framework integrates principal component analysis for dimensionality reduction, mutual information-based feature ranking for relevance assessment, and recursive feature elimination for optimal feature subset selection. The optimized feature space is subsequently used to train and evaluate multiple ML classifiers, including logistic regression, support vector machine, random forest, gradient boosting, and extreme gradient boosting (XGBoost). Experimental evaluation demonstrates that the proposed framework effectively reduces the original feature set from 13 to 6, achieving a feature reduction rate of 53.85% while preserving critical diagnostic information. Among the evaluated classifiers, XGBoost achieved the highest performance with an accuracy of 98.31%, precision of 98.06%, recall of 97.95%, F1-score of 98.00%, and ROC-AUC of 99.02%. Comparative analysis with recent state-of-the-art CVD prediction methods confirms the superiority of the proposed HFSDR-CVD framework in terms of predictive accuracy, feature optimization, and computational efficiency.
Abstract The concept of a smart campus (SC) related to sustainable development has attracted considerable research interests, yet a clear model with key components is still lacking in Vietnam. This study seeks to identify investment priorities for transforming Vinh University (VU) into a smart and sustainable campus. Using Cronbach's Alpha, we assessed the reliability of six dimensions based on 35 indicators drawn from literature, the university's context, and experts' perspectives before conducting an online survey to collect feedback from stakeholders, including students, faculty, staff, alumni, and employers. Importance-performance analysis (IPA) highlighted VU's strong performance indicators, such as smart governance (SG)9 (3.17), smart services (SS)4 (3.10), and smart education (SE)3 (3.07). Although expected targets for prioritized improvement included SE7 (3.39), SG9 (3.42), and SG3 (3.36), the low I-P gap and specific conditions at VU emphasize the need for VU to enhance four priority sections, including promotion, strategies (plan, funding, and infrastructure), collaboration, and sustainability, to further smart development.
Adaptive antenna arrays are essential in defense, radar, and secure communications, where accurate reception of desired signals is required despite interference and noise. Real-world environments contain multiple jammers, dynamic interference, moving sources, and time-varying noise, which reduce the effectiveness of conventional beamforming and null-steering techniques. To address these challenges, this work analyzes an adaptive array pattern-nulling technique based on Modified Invasive Weed Optimization with Laplace distribution (MLIWO). After signal acquisition and digitization, high-resolution algorithms such as multiple signal classification (MUSIC) and estimation of signal parameters via rotational invariance techniques (ESPRIT) are used to estimate interference directions and spatial characteristics. MLIWO then iteratively optimizes the array weight vector to preserve main-lobe fidelity while producing deep nulls in interference directions. The optimized weights form the desired radiation pattern, evaluated using null depth (ND), side-lobe level (SLL), array gain, and signal to interference plus noise ratio (SINR). MATLAB results show that MLIWO achieves significant interference suppression, obtaining ND values of -80.75 dB at -10 degrees and -88.87 dB at -42 degrees, outperforming IWO and crow search optimization (CSO).
Timely and precise detection of adolescent stress is critical for early intervention, yet current CNN- or Transformer-based models fail to jointly capture both structural facial dependencies and temporal emotion dynamics. To address this gap, we propose a novel spatiotemporal self-attentive graph-TCN (STG-TCN) framework that integrates graph-based spatial reasoning with temporal convolutional modeling. Unlike conventional approaches that concatenate spatial and temporal features, our method employs a cross-attention fusion mechanism to align and enrich these modalities. The proposed model was evaluated on the dataset for affective states in e-environments (DAiSEE) under strict cross-subject protocols, achieving 90.2% classification accuracy, a 12% improvement over graft attention network (GAT)-only baselines, and a 9% improvement over temporal convolutional networks (TCN)-only baselines. Furthermore, STGTCN reduced false negatives by 20%, demonstrating high sensitivity to subtle stress cues such as micro-expressions and transient muscle contractions. These findings establish STG-TCN as the first framework specifically tailored for adolescent stress recognition in real-world video settings, thereby advancing the field of affective computing with both methodological innovation and practical implications for unobtrusive mental health monitoring.
Abstract Purpose This paper comprehensively reviews the effects of machine learning (ML) on the diagnosis and treatment of four different diseases: electrocardiogram (ECG), diabetes, chronic kidney disease (CKD), and breast cancer. The primary objective of this paper is to investigate how ML algorithms address challenges related to early detection, diagnostic precision, and data security across various medical fields. Method The study employs a systematic review approach, formulating research questions and extensively analyzing existing literature. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology is utilized in the literature on these four diseases. This analysis focused on ML applications in performance metrics, disease prediction, and advancements in the field. The review integrates bibliometrics, PRISMA-based study selection, and an ML technique-based in-depth comparative review. Results The review indicates the transformative impact of ML on disease detection and management. Breast cancer research represents the most significant growth and citation rates, indicating its high relevance. CKD and diabetes research also demonstrated notable advancements but faced recent declines in output. However, electrocardiogram research requires further innovation with advancements in other areas. Conclusion ML algorithms have great potential to solve data security, early detection capabilities, and diagnostic accuracy problems, and provide new solutions in various medical fields. Future research should focus on improving healthcare delivery and patient outcomes by enhancing model robustness and integrating emerging technologies.
Abstract Videonystagmography (VNG) is a diagnostic technique that utilizes infrared video goggles to record eye movements and assess the vestibulo-ocular reflex (VOR) by tracking involuntary eye movements such as nystagmus and saccades. In recent years, eye-tracking technologies have gained attention in educational research for understanding learner engagement and cognitive processing. The novelty of this study lies in the application of clinical-grade VNG technology to quantitatively correlate saccadic eye movement parameters with academic performance in engineering students, an area that remains largely unexplored. An investigation was carried out to analyze the relationship between VNG parameters and the academic performance of undergraduate engineering students. A total of 36 students (18 males and 18 females), aged 19–20 years, participated in the study. Performance scores were computed as the average of six theory subjects. The VNG parameters analyzed included velocity, precision, and latency of horizontal and vertical saccades for both eyes. Statistical analysis revealed that saccadic velocity, precision, and latency are mutually independent parameters, while strong correlations exist between the left and right eyes. Regression analysis showed that vertical saccadic latencies of both eyes significantly predict academic performance, while horizontal latency of the right eye also showed statistical significance. Gender was found to significantly influence most eye-tracking parameters. Although conducted in a controlled laboratory environment with a limited sample size, this study demonstrates the potential of VNG-based eye-tracking metrics as objective indicators of cognitive processing in engineering education, offering new directions for technology-enabled educational assessment.
Diagnosis of prostate cancer is an area of medical research of critical importance, in which advancements in imaging technologies have much improved detection as well as treatment outcomes. Although substantial progress has been made concerning the application of machine learning (ML) and deep learning (DL) techniques, few systematic reviews have examined these techniques in the context of multiparametric MRI (mpMRI) and diffusion-weighted synthetic imaging (DWSI). Existing studies often focus on individual methods or imaging modalities, leaving a gap in understanding how these techniques are integrated and optimized for diagnostic precision. This motivated this review paper to comprehensively review and summarize the new automated methods for prostate cancer diagnosis, particularly through the use of mpMRI and DWSI imaging. It explores various imaging modalities and their integration with DL and ML techniques to improve diagnostic accuracy. The review assesses the effectiveness of these advanced imaging approaches in Gleason score (GS) estimation and highlights the challenges associated with each modality. The review systematically compares performance evaluated by specific feature values such as specificity, F-measure, precision, and accuracy of several ML and DL algorithms for prostate cancer diagnosis. Alongside this, the review brings to attention the current limitations of the approaches and points to future research directions with an emphasis on the innovative requirement for finding better generalization methods to mitigate diagnosis problems in prostate cancer management.
Buying intent detection is a central challenge that cuts across several artificial intelligence domains, including natural language processing, knowledge representation, and decision-making systems. Inferring a user’s purchase intent correctly is of critical importance in real-world applications such as e-commerce, customer relationship management, and personalized recommendation systems, where it directly impacts conversion rates, user engagement, and operational efficiency. Over the years, a wide range of approaches have been proposed that span deep-learning models, knowledge-graph based reasoning, reinforcement learning, and more recently agent-based and multi-agent systems. This review systematically analyzes the evolution and efficiency of these paradigms with an eye toward real-world applicability. Deep-learning models based on transformers are among the most powerful detection models, given large labeled datasets, but offer little in terms of interpretability or robustness under domain shift. Methods leveraging knowledge graphs augment reasoning and interpretability by modeling structured relationships but come at significant construction and maintenance costs. Reinforcement learning introduces the ability to learn adaptively in sequential and dynamic environments but is sensitive to reward design and sample efficiency constraints. AI-agent systems provide more comprehensive autonomy and facilitate multi-step coordination tasks while introducing novel challenges related to reliability, latency, and operational overhead. We provide a unified taxonomy that synthesizes these approaches and compare them on several dimensions, including detection accuracy, quality of the induced reasoning, robustness to domain changes, and efficiency in the deployment setting. Unlike previous surveys, this work puts an emphasis on deployment-realistic benchmarks such as latency, cost of inference, quality of the grounded output, and tool success rate, supported by reported benchmarks and ablation studies. We conclude by pointing out open challenges regarding scalability, interpretability, and integration of symbolic and data-driven methods, giving practical insights for both researchers and practitioners while developing intent-aware AI systems.
Diabetic retinopathy (DR) is the leading cause of blindness worldwide and refers to progressive degeneration of the retina. Early diagnosis and treatment of this condition are important to prevent permanent damage. Lipid deposits in the retina, which are found as exudates, are one of the major indicators of DR. The accuracy of detection, however, is affected by several factors, such as noise and image resolution. Deep learning (DL) models require significant computational power and vast quantities of data, whereas traditional machine learning methods are not capable of dealing with these challenges. This study proposes a fusion of machine learning and DL techniques for the automation of exudate segmentation and classification. The method applies preprocessing, which reduces noise and enhances input quality. A convolutional neural network (CNN) with an encoder-decoder structure based on U-Net performs pixel-wise exudate segmentation with precise localization. We applied a DL technique with pretrained DenseNet models to classify the exudates as mild, moderate, and severe. Using techniques such as rotation, zooming, and flipping, we enhance the model's capability and help with data shortages. The standard datasets of digital retinal images for vessel extraction (DRIVE), structured analysis of the retina (STARE), diabetic retinopathy database (DIARETDB1), and IDRiD were used to evaluate the model. The results of the study show segmentation accuracy, measured by dice, of 0.91 (training), 0.89 (test), and 0.88 (validation). Furthermore, the classification shows an accuracy of 0.92 (training), 0.90 (test), and 0.88 (validation). Finally, both sensitivity and specificity were above 85%. The combination of preprocessing and transfer learning reduces the computational burden and makes it clinically usable in real-time.
Orthogonal frequency division multiplexing (OFDM) is the most sought choice when it comes to high speed applications owing to its superlative performance in fading environment. However spectral efficiency of OFDM can be further enhanced by using OFDM with index modulation (OFDM-IM) where index of the active subcarrier is also used to carry the payload along with constellation vector. Similar to its predecessor, the OFDM-IM system also suffers from high peak-to-average power ratio. This paper uses discrete wavelet transform (DWT) based OFDM-IM system where the inverse fast Fourier transform (IFFT) block is replaced with DWT block and results show better performance of DWT based systems over IFFT based systems in terms of Peak-to-Average Power Ratio (PAPR) reduction. Besides this we also evaluate the performance of system over alpha - kappa - & micro; fading channel. alpha - kappa - & micro; is one of the most sophisticated fading channels which fits very closely to practical data statistics.
The main reason for ocular deficiency is diabetic retinopathy, which is prevalent among people aged 25-74, significantly affecting health care and socioeconomic systems. Early detection can prevent vision loss in nearly 90% of cases, but retinal fundus images often suffer from noise and poor illumination, limiting automated analysis. This study proposes an integrated image enhancement and classification framework using Lab color-space enhancement, Wiener filtering, adaptive fuzzy Tsallis entropy segmentation, and curvelet-based feature extraction. The proposed color dominance and boosted Remora optimization algorithm with deep adversarial approach achieved high accuracy, precision, sensitivity, and robustness on a retinal fundus dataset.
Biometric-based access control systems are gaining popularity over traditional security methods such as keys orpersonal identification numbers (PINs), offering stronger protection for homes and organizations as a safer and smarter alternative. This research presents a sequential fingerprint authentication framework in which multiple biometric inputs must be verified in a predefined order to grant access. Unlike conventional single-fingerprint authentication systems, the proposed design transforms biometric authentication into a biometric passcode mechanism, requiring both correct identity and correct order of fingerprints. This additional layer significantly increases resistance to unauthorized access attempts while maintaining a low-cost embedded implementation suitable for practical access control applications. This solution is flexible and affordable, and it can be used in residential, commercial and industrial environments. By combining fingerprint authentication with sequential verification, the proposed system provides a robust, scalable and user-friendly security solution. The integration with Arduino and relay control further enhances its flexibility, making it a practical choice for modern smart security systems.
