Browse Items (16481 total)
Sort by:
-
Impact of Social Media on Patience, Anxiety and Stress
Social media has become deeply integrated into everyday life, constantly connecting us to various platforms. While people were once believed to be shaped by those they spent time with physically, the content we consume and individuals we follow online now significantly influence our thoughts, emotions, and behaviours. Despite growing concerns about these effects, limited research has examined the relationship between social media usage patterns and psychological well-being among young adults. This study examined whether social media usage patterns affect patience, stress, and anxiety among young adults. A cross-sectional survey was conducted with 124 participants aged 18-30 years. The survey collected demographic information and assessed social media usage patterns, with scores computed for patience, anxiety, and stress. Results indicated that stress levels were positively correlated with social media usage (r=0.33, p< 0.001), with Instagram emerging as the most widely used platform. A paired t-test revealed that participants significantly underestimated their actual screen time compared to their perceived screen time, t(123) = [insert your t-value], p=0.0002. Correlation analysis also indicated that screen time was positively associated with increased anxiety (r=0.22, p=0.014) and negatively associated with patience (r=-0.19, p=0.035). These findings highlight the importance of developing digital self-awareness, encouraging individuals to maintain control over their social media usage rather than letting these platforms dictate their levels of stress, patience, and anxiety. 2025 IEEE. -
Comparative Performance Evaluation of GEO, MEO, and LEO Satellite Networks under Traffic Attacks
This paper presents a comparative evaluation of geostationary (GEO), medium-earth orbit (MEO), and low-earth orbit (LEO) satellite constellations under realistic traffic attack models. We use OMNeT++ v6.1 with the INET-4.5 framework for simulation and Python for analysis. Key performance metrics include end-to-end latency, throughput, packet delivery ratio (PDR), and resource utilization measured under normal and attack conditions. Our results indicate that MEO yields the highest throughput and resource utilization, while LEO offers the lowest latency. We provide a clear description of the simulation conditions, attack models, and statistical methods used to evaluate resilience under degraded operation. 2025 IEEE. -
Privacy Preserving Authentication Using Zero Knowledge Proofs
Conventional authentication techniques, such as one-time passwords and passwords, are extremely susceptible to data breaches, credential theft, and phishing attacks. These vulnerabilities are increased when using shared or public devices. This paper proposes a password-less authentication architecture for various environments and organization based on Zero-Knowledge Proofs in order to overcome these issues. The proposed model ensures that no sensitive credentials are sent or retained by having a user demonstrate that they possess a secret without disclosing it to the server. In doing so, the attack surface linked to traditional login methods is greatly reduced. The framework is meant to be scalable, lightweight and easy to integrate with learning management systems, corporate sites, online test platforms, and university websites. 2025 IEEE. -
Analyzing Deep Learning Architectures in Cotton Crop for Precision Disease Diagnosis
Cotton is an important cash crops worldwide, providing raw materials for the textile industry and is the basis of livelihood of millions of farmers. In India, it has an important place in the agricultural economy, which contributes significantly to both domestic consumption and export income. However, cotton production is highly sensitive to infection of various diseases and insects, such as bacterial scorching, powdery mildew and targeted spots, which can cause severe yield reduction and economic loss. Traditional disease management methods often depend on manual inspection, which is difficult to scale in time consuming, human error and large cultivated areas. Therefore, it is necessary to detect the initial and accurate detection of the disease to ensure plant health and maximize productivity. This study examines advanced intensive teaching methods for automatic cotton disease diagnosis, and compare the performance of VGG16 and ResNet18 architecture. Experimental results showed that the VGG16 model achieved verification accuracy of 99.69%, while ResNet18 achieved an accuracy of 99.58%. In addition, a real time forecasting interface was developed from the URL provided by the user to classify images of cotton leaves, making practical signs possible for use in the area. This research highlights effectiveness of deep learning in improving accurate agriculture, which helps in timely detection of diseases to reduce the loss of crops. 2025 IEEE. -
Emotion Trajectory Analysis and Model Comparison for Hate Speech and Radicalization Detection in Code-Mixed Platforms
The growing presence of multilingual and codemixed content on social media creates major challenges for automated emotion recognition and mental health support. In this work, we introduce an emotion-aware computational framework that processes code-mixed Indian language comments and predicts user emotions with high accuracy, followed by context-aware support suggestions. Our dataset comes from the AI4Bharat IndicNLP corpus [14] and the Dravidian-CodeMix sentiment dataset [15], featuring a variety of multilingual user comments. To maintain linguistic consistency, we translate the raw texts into English using Google Translator and then preprocess them through normalization, tokenization, and stopword removal. We use three advanced transformer-based models, DistilBERT (six emotions), DistilRoBERTa (seven emotions), and RoBERTa GoEmotions (27+ emotions), to categorize the emotions in the comments. We compare predictions across the models and select the most reliable label for each text, which is further verified through manual checks with human annotators. This process results in a curated dataset labeled with emotions and enriched with model provenance. With this dataset, we train a Logistic Regression classifier using TF-IDF features to create an efficient, explainable prediction pipeline. The system classifies emotions and provides tailored suggestions based on emotional states, improving user support in online interactions. Experimental results show the robustness of the pipeline and its ability to adapt to various code-mixed inputs. This study offers an integrated dataset-model-suggestion framework that advances emotion recognition in multilingual contexts and supports the creation of practical emotion-aware digital systems. 2025 IEEE. -
Inclusive Interfaces: Enhancing Voice Assistants for Diverse Speech Abilities
The main goal of this research is to highlight the considerable challenges that people suffering from dysarthria have to face while using Automatic Speech Recognition (ASR) technologies and virtual assistants. Current voice-controlled technologies struggle to understand disordered speech, which reveals a significant gap in research about how to adapt these technologies to speech changes caused by health conditions. We propose to explore non-verbal voice signals as a quieter potential solution. The work uses a systematic literature review and comparative analysis across various scenarios to point out the weaknesses of ASR in terms of accuracy and security. The research proves that non-verbal input always provides greater performance in accuracy, reliability, and convenience. This analysis lays the groundwork for designing more accessible technologies. Designed for researchers and developers, this research is an avenue for future work where future efforts address the empirical evaluation of a specialized non-verbal system to support user autonomy. 2025 IEEE. -
Fairness-Aware and Interpretable Depression Detection on Social Media Using BERT with Gender Bias Mitigations
Reddit and similar social media platforms offer substantial information regarding mental health issues. The automatic detection of depression raises different issues pertaining to fairness and transparency. This paper presents a Fairness Aware and Interpretable Depression Detection framework that utilizes BERT and incorporates an explicit gender bias mitigation mechanism. Data were obtained from gender-specific forums on Reddit. The Mistral language model based classifier was used to set a high confidence threshold, which helped in inferring gender labels while both depressed and non-depressed were among the patients assigned the labels. A balanced dataset with four groups (Depressed-Male, Depressed-Female, NonDepressed-Male, NonDepressed-Female) was prepared. Two pipelines were carried out where one involved a baseline BERT classifier while the other employed a fairness aware BERT model that incorporated gender embeddings during the training phase. The models were assessed using accuracy, precision, recall, F1 score, and confusion matrices and the fairness metrics applied were Demographic Parity Difference (DPD) and Equal Opportunity Difference (EOD). To enhance the model's reasoning transparency, SHAP was applied due to its capability to provide clear and comprehensive explanations. The results indicated that the fairness centered model effectively reduced gender biasness and equalized error rates among the different groups without losing its original accuracy. The essential point is that the model had learned to give precedence to clinical indicators over gender specific language. This study suggests a roadmap for the creation of ethical AI by combining fairness, interpretability and high performance into a seamless framework. 2025 IEEE. -
Audio Recognition of Animals Using Optimized Deep Learning Techniques for the Conservation of Wildlife
The classification of animal sounds has emerged as a vital tool in contemporary research, offering numerous benefits for animal occurrence records, taxonomic research, and behavioral studies. However, the problem of accurately identifying animal species based on their vocalizations remains a significant challenge, particularly in real-world environments where background noise and variability in sound patterns can hinder classification accuracy. In this paper addressed this challenge by proposing a CNN-optimized approach for classifying animal sounds. In order to enhance the number of sound samples, utilized augmentation techniques to extract animal sounds from the Kaggle animal sounds dataset. The animal sounds totally 600 audio samples are used. To improve performance, this model was developed using feature extractions from the MFCC, ZCR, and Mel-Spectrogram. The seamless deployment of forest department workers is ensured by the interpretability of our model for real-world applications related to wildlife conservation and monitoring. The main goal is to successfully identify animals using auditory properties, such as tiger, leopard, elephant, and otter noises, based on their vocalizations. Additionally, The optimized CNN and LSTM for sound classification. The Optimized CNN outperformed all other models, achieving an outstanding 98.32 % training accuracy rate. 2025 IEEE. -
A Mathematical Modelling of Macronutrient Composition on Energy Content in Indian Breakfast Foods
The relationship between macronutrient composition and energy content plays a crucial role in dietary planning and metabolic health. Regression analysis and Anova are used in this study to analyze the effects of macronutrients (fats, sugars, and proteins) and micronutrients (iron) on the calorie content of Indian breakfast items. This study develops a data driven explanation of the primary determinants of caloric density by examining a dataset that include nutritional characteristics and environmental impact components, such as carbon footprint. The study also looks at how changes in the combination of macronutrients impact the energy balance of the diet, providing information on how to select foods that are best for sustainability and health. Considering environmental factors into the analysis also emphasizes the necessity of sustainable food consumption habits. The results contribute in the development of individualized nutrition plans, supporting people in making wise food choices with the least possible negative environmental effects. This study provides the groundwork for further research in the food industry and public health by highlighting the importance of establishing a balance between sustainability objectives and nutritional sufficiency. 2025 IEEE. -
Transforming Classrooms with Adaptive Content: From One-Size-Fits-All to Personalized Mastery
Educational disparities continue to be a great hindrance in contemporary learning settings as a result of the inability of conventional methods, founded on the single-size-fits-all principle, to address the divergent demands of students' thinking. This envisioned technological intervention is a paradigm shift to personalized delivery of educational content through cutting-edge artificial intelligence methodologies. The strategy bridges the gap between innovation in technology and pedagogical impact, resulting in democratization of learning experience. This is especially transformative for students from marginalized communities, who have in the past been excluded from accessing customized educational resources because of the lack of access to quality education. 2025 IEEE. -
Post-Quantum Cryptography for Securing Next-Generation Communication Networks
Advancements in Quantum Key Distribution (QKD) and lattice-based encryption are paving the way for PQC adoption, but challenges remain, such as performance overhead and compatibility with existing infrastructure. It evaluates whether PQC schemes are feasible for real-time applications in high-speed, low-latency networks and analyzes the security-performance trade-offs. We investigate standardized candidates from NIST's PQC Project (e.g., CRYSTALS-Kyber, Dilithium) and their resistance to hybrid attacks. In addition, we also investigate the hardware acceleration (e.g., FPGA, ASIC) approach to alleviate the latency bottleneck. Transition strategies, such as hybrid cryptography (the coupling of classical and PQC algorithms) and zero-trust frameworks to maintain backward compatibility, are a key focus here. We further discuss side-channel vulnerabilities specific to PQC implementations and suggest mitigation strategies. These findings emphasize the need for a continued focus in areas such as scalability, standardization and quantum secure key distribution and the importance of collaboration between academia, industry and policymakers."By tackling these issues, PQC can secure next-gen networks from quantum dangers while aging to be efficient and trustworthy. 2025 IEEE. -
Ultrafast nonreciprocal transmission modulation in metasurfaces with epsilon-near-zero materials
Nonreciprocity refers to the difference in received to transmitted ratio when the source and detector are interchanged [1]. Optical isolator - component which allows transmission in one direction - is a canonical example of a nonreciprocal device. Nonreciprocity can be achieved through three known pathways; (i) materials with asymmetric permittivity or permeability tensors, such as ferrites; (ii) nonlinear light-matter interactions[2-4]; and (iii) time-varying systems[5]. While traditionally nonreciprocal components are quite large in size, nanofabrication of metasurfaces has enabled their miniaturisation to the nanoscale. However, ultrafast nonreciprocal responses at the nanoscale remain still a challenge. Here we design and study metasurface with an epsilon near zero material indium tin oxide (ITO) that enables ultrafast switching of refractive index via Kerr nonlinearity, in order to achieve optical isolation. 2025 IEEE. -
HRL-ViT: Human-Robot Collaborative Vision Transformer for AIoT-Enabled Leaf Disease Detection in Precision Agriculture
The combination of artificial intelligence and Internet of Things (AIoT) technologies is changing precision agriculture by making it possible to automatically check the health of crops. Early detection of leaf diseases is still important for stopping yield losses, but regular convolutional neural networks (CNNs) often don't work as well when they have to deal with different textures, lighting changes, and noise on the field level. To address these constraints, this study presents HRL-ViT, a Human-Robot Collaborative Learning framework that utilizes Vision Transformers for leaf disease identification. The frame-work merges the global attention feature of Vision Transformers with a human-in-the-loop approach, wherein predictions with low confidence are validated by experts and used to improve the model over time. The system is also made for edge-based AIoT deployment, which lets you analyze data in real time in agricultural settings. Experimental research utilizing both benchmark datasets and field-acquired images demonstrates that HRL-ViT consistently surpasses baseline CNN and Transformer models, attaining superior accuracy, precision, and recall while minimizing false detections. Transformers' attention maps can be visualized to make them even easier to understand, which helps users trust them and make decisions. In general, HRL-ViT shows a lot of promise for use in autonomous robotic platforms. It offers an explainable and scalable way to find diseases in precision agriculture. 2025 IEEE. -
Harnessing the Internet of Things for Sustainable Urbanization: A Framework for Resilient Smart Cities
The Internet of Things (IoT) represents a revolutionary technology, which facilitates the emergence of dynamic network of inter-connected devices that can flawlessly communicate and share information. It also merges various technologies, hardware systems, and software frameworks to form an all-embracive ecosystem, which comprises data, people, devices, and smart communication. In a country like India, which exhibits major regional differences in the degree of technology and IoT adoption, it can easily be considered that the idea of IoT has the enormous potential to transform the urban development process and introduce it into the paradigm of improvability, cost-efficient, and maintenance friendly solutions. The issues of smart cities are central to the development of a country and better life of citizens. In India, use of IoT in smart city projects makes it easier to provide core services to constituents, implement transparent governance strategies, and build sustainable urbanization. With the help of IoT, different systems may correlate effectively which will provide a reliable access to the data and the establishment of new opportunities in the form of innovative digital services that would meet the requirements of city residents. The present research paper explores the centrality of IoT in the development of smart cities in India. It discusses the policy framework of IoT in the country, outlines major drivers of and benefits of IoT-based smart city solutions, analyzes consumer preferences and demographics. With this comprehension, the paper would like to suggest practical insights concerning the use of IoT in creating inclusive and technology-driven cities of India. 2025 IEEE. -
A Comprehensive Comparison of MobileNet, ResNet50 and InceptionV3 for Efficient Plant Pathology Detection
Plant diseases have a strong impact on agricultural productivity due to economic factors and a reduction in crop quality. This work focuses on the classification of apple leaves into four classes: healthy, rust-infected, scab-infected, and infected by both diseases, using the Plant Pathology FGVC7-2020 dataset that contains 3,642 images in total. The work involves the analysis of three sophisticated deep learning architectures: ResNet50, MobileNet and InceptionV3.It turned out that MobileN et showed the highest performance, reaching a 92% accuracy rate; it was followed by ResNet50 with 75% accuracy and InceptionV3 at 73%, hence underlining its better generalization capability and efficiency in classifying. We discuss the proposed methodology, which includes data preprocessing techniques, experimental results and final conclusions, is discussed in detail. These results underline the fundamental importance of determining an appropriate neural network architecture for the recognition of plant diseases, which is of prime importance to improve agricultural productivity. 2025 IEEE. -
Implementing Machine Learning for Early Detection and Prognostic Modeling of Chronic Diseases
The employ of deep learning methods for the diagnosis and prognosis model of chronic diseases is an important discovery to change the healthcare service. Some of the chronic diseases which prevalence and incidence rates remain high globally include diabetes, cardiovascular diseases, chronic kidney diseases, and cancers. There is nothing more critical than early diagnosis and accurate prediction of the patients' condition and the best course of action that has to be taken. This paper aims at examining the possibility of utilizing ANN, Random Forest, XGBoost, and CNN to forecast the occurrence of the. Due to integration of big and varied data which involve clinical characteristics, biochemical parameters and medical images among others, ML models have the ability recognize complex relations not easily recognizable by conventional diagnostic procedures. These illustrations prove that deep learning models or more specifically the convolutional neural networks for image diagnosis outperform other traditional methods in performance and prognosis. Nevertheless, some issues, such as data quality, model's interpretability, and its implementation into clinical practice, are still present. The challenges appeared in this paper are key to understanding the future of ML in healthcare as they can pave the way to the integration of such models into practice, therefore leading to early detection, better prognosis, and effective management of chronic diseases. This paper aims at exploring on how ML can be of significance in transformation of the health care sector and orderly improve patients care. 2025 IEEE. -
A Systematic Review and Modern Approaches for Bio Signal Based BCI's
Bio-signals play a very critical role in modern medicine, especially for paralyzed patients. The development of Brain-Computer Interface (BCI) systems, which allows direct brain-to-external device contact, is made possible by these signals, creating new ways for medical intervention and rehabilitation. Using these bio-signals, modern medicine has made great strides toward developing intelligent devices that enhance the quality of life for people who are paralyzed. These include improved mobility, supported communication devices, and environmental control. In this survey, we thoroughly assess the latest BCI-based smart devices and their medical applications. By finding and examining current methods, technologies, and practical uses, we seek to showcase the effectiveness and potential of brain-computer interfaces (BCIs) in providing new ways to treat paralysis. Currently, bio-signal-based control systems have been continuously used rapidly in biomedical devices and assistive robots to enhance and improve the quality of life for disabled and elderly individuals. Among these, electromyograph (EMG), electroencephalography (EEG), and electrooculography (EOG) bio-signals are highly used for improving modern technologies. The primary objectives of this paper are to detail the techniques used in Brain-computer interface-based smart devices and their applications, used in healthcare. Additionally, this paper provides an overview of applications controlled through these bio-signals and discusses the research challenges in developing these control systems. 2025 IEEE. -
JivaCare: A Smart Health Care Application Integrating Home Remedies for Holistic Well-Being
JivaCare is a smart health care application that aims to diagnose common illnesses and recommend home remedies by leveraging advancements in Artificial Intelligence (AI) and Machine Learning (ML). Initially, a custom dataset of diseases and symptoms was used to develop a disease-diagnoses chatbot. However, challenges such as limited accuracy and rigid disease-symptom mapping were identified. To address these, GPT-3.5 API was integrated via RapidAPI, enhancing conversational quality without requiring extensive training while improving response quality. A MySQL database was implemented to store conversational history and session-based memory. Subsequently, the focus transitioned to a text classification approach using a 5,634-sample dataset from HuggingFace. This enabled flexible symptom-to-disease classification, overcoming the limitations of the initial dataset. Five machine learning models were evaluated, with Logistic Regression achieving the highest accuracie of 85 % after fine-tuning its hyperparameters. Additionally, neural network architectures such as GRU, RNN and CNN were also employed, achieving validation accuracies of 76%, 82% and 84%, respectively. The results demonstrate the effectiveness of integrating ML and deep learning techniques for accurate disease prediction and remedy recommendation. This work can establish the foundation for a scalable and user-friendly healthcare system, bridging the gap between AI and personalised natural health benefits. 2025 IEEE. -
JivaCare: A Smart Health Care Application Integrating Home Remedies for Holistic Well-Being
JivaCare is a smart health care application that aims to diagnose common illnesses and recommend home remedies by leveraging advancements in Artificial Intelligence (AI) and Machine Learning (ML). Initially, a custom dataset of diseases and symptoms was used to develop a disease-diagnoses chatbot. However, challenges such as limited accuracy and rigid disease-symptom mapping were identified. To address these, GPT-3.5 API was integrated via RapidAPI, enhancing conversational quality without requiring extensive training while improving response quality. A MySQL database was implemented to store conversational history and session-based memory. Subsequently, the focus transitioned to a text classification approach using a 5,634-sample dataset from HuggingFace. This enabled flexible symptom-to-disease classification, overcoming the limitations of the initial dataset. Five machine learning models were evaluated, with Logistic Regression achieving the highest accuracie of 85 % after fine-tuning its hyperparameters. Additionally, neural network architectures such as GRU, RNN and CNN were also employed, achieving validation accuracies of 76%, 82% and 84%, respectively. The results demonstrate the effectiveness of integrating ML and deep learning techniques for accurate disease prediction and remedy recommendation. This work can establish the foundation for a scalable and user-friendly healthcare system, bridging the gap between AI and personalised natural health benefits. 2025 IEEE. -
Dynamic Load Balancing on Switches of Software Defined Network Managed by OpenDayLight Controller
In recent times, the world is becoming a global village where connectivity is a new norm irrespective of geographical location. Within corporate networks, huge setbacks are faced due to a lack of efficient resource management. Load balancing is inevitable to cater to a reliable, faster, and congestion-free communication experience for exponentially increasing online enterprises. Dynamic network resource management for high performance and low data transmission latency in a network is necessary. The major issue faced by the traditional network is that it relies on static hardware switches. Software-Defined Network approaches paved a way to overcome the limitations of traditional networks. This research proposes the Dynamic Load Balancing Algorithm for Software-Defined Networks to utilize network resources optimally. The major function of the proposed algorithm is to determine alternative paths and further distribute the incoming and outgoing network flows to achieve optimum network resource utilization with faster traffic flow completion. The experiment is performed with the OpenDayLight Controller on the Mininet simulator, which emulates the network with the novel scheme. The results prove that the proposed solution has accomplished the benchmarks of optimum throughput, reduced redundancy, and reduced flow completion time. 2025 IEEE.