Browse Items (16481 total)
Sort by:
-
A machine learning model to predict suicidal tendencies in students
[No abstract available] -
Mul-Sensis: Multilingual Sentiment Analysis Framework for Emotion Detection
Sentiment analysis is a pivotal Natural Language Processing (NLP) task that enables the extraction of actionable insights from textual data, particularly social media. With the rise of public discourse on platforms like Twitter, analyzing sentiment trends has become crucial for decision-making in domains such as policy implementation, feedback evaluation, and public opinion monitoring. Mul-Sensis employs a hybrid approach combining transformer-based models with classical machine learning algorithms to enhance sentiment classification. The system integrates advanced preprocessing techniques to address linguistic complexities like sarcasm, idiomatic expressions, and domain-specific nuances. A robust hybrid annotation approach, incorporating both human expertise and machine-assisted methods, ensures high-quality, bias-free sentiment labeling. This study contributes a scalable, interpretable, and domain-agnostic framework for sentiment analysis, offering valuable insights for policymakers, researchers, and industries relying on textual data analytics. The findings highlight the transformative potential of hybrid and ensemble-based NLP approaches for understanding public sentiment across diverse cultural and linguistic contexts. The Author(s), under exclusive license to Springer Nature Switzerland AG 2026. -
A Hybrid Deep Learning Model Using U-Net and Vision Transformer for Artificial Intelligence Powered Cervical Stenosis Diagnosis
This study presents a deep learning-based approach for the classification of cervical stenosis using MRI spine images, integrating multiple phases such as preprocessing, segmentation, feature extraction, and classification. A U-Net-based segmentation model effectively delineates key anatomical structures, including the spinal canal, intervertebral discs (IVDs), and neural foramen, improving feature extraction and classification accuracy. Furthermore, ResNet-50 is employed for feature map generation, leveraging deep hierarchical representations to extract meaningful spatial patterns from MRI slices. For classification, a Vision Transformer (ViT)-based model is utilized, taking advantage of its self-attention mechanism to capture both local and global dependencies within MRI images. Unlike conventional CNN-based models, ViT processes MRI scans as patches, enabling a more context-aware analysis of stenotic regions. The model is trained using an 80%20% train-test split and evaluated using standard performance metrics, achieving an accuracy of 92.60%, precision of 90.16%, recall of 95.43%, and an F1-score of 91.56%. These results indicate that the ViT model outperforms traditional CNN-based classifiers in cervical stenosis detection, ensuring higher sensitivity and specificity in real-world clinical applications. The Author(s), under exclusive license to Springer Nature Switzerland AG 2026. -
Artificial Intelligence-Driven Lumbar Stenosis Diagnosis: A Deep Learning Pipeline for MRI-Based Segmentation and Classification
Lumbar spinal stenosis is a prevalent musculoskeletal disorder that requires accurate diagnosis through magnetic resonance imaging (MRI). However, manual interpretation of MRI images is time-consuming and subject to inter-observer variability. This study proposes an automated deep learning-based pipeline for lumbar stenosis identification, integrating advanced methodologies for preprocessing, segmentation, feature extraction, and classification. The pipeline consists of Super-Resolution Convolutional Neural Network (SRCNN) for MRI image enhancement, SegNet for segmentation of the spinal canal, intervertebral discs (IVDs), and neural foramen, Convolutional Block Attention Module (CBAM) for feature refinement, and Swin Transformer for final classification. The proposed method was evaluated on a publicly available multicenter lumbar spine MRI dataset, comprising 218 patient studies with 447 MRI series. Model performance was assessed using accuracy, recall, precision, and F1-score, achieving 95.2% accuracy, 89.82% recall, 92.3% precision, and an F1-score of 96.12%. The results demonstrate that SRCNN enhances MRI quality for improved segmentation, CBAM strengthens feature extraction, and Swin Transformer effectively classifies stenosis cases. This study highlights the efficacy of AI-driven methodologies in lumbar spine MRI analysis, offering a potential computer-aided diagnosis (CAD) tool for clinical applications. Future work may focus on optimizing model efficiency and improving generalization across diverse imaging protocols. 2025 IEEE. -
Precise cervical cancer cell boundary denoising and segmentation with adaptive wavelet-spectral enhancement
Accurate segmentation of cell nuclei in cervical cytology images is crucial for automated cervical cancer screening, yet existing methods struggle with blurred boundaries, noise-induced degradation, and topologically implausible predictions. The current research proposes Cell-Seg Tool, a novel triplet-branch diffusion AI tool that synergistically integrates three innovations to address these limitations. The Wavelet-Enhanced Contour Refinement Branch employs a learnable multi-scale discrete wavelet transform with adaptive coefficient attention to dynamically enhance boundary features across horizontal, vertical, and diagonal orientations. The Adaptive Spectral Noise Suppression module performs dual-domain processing using DCT-based filtering and uncertainty-guided fusion, coupled with bidirectional anchor semantic feedback to couple cross-branch information. The Topology-Aware Hybrid Loss integrates a focal Tversky loss, a persistent homology loss, a directional boundary loss, a skeleton completeness loss, and a diffusion-noise MSE loss for multi-objective optimization. Comprehensive experiments on multiple datasets demonstrate superior performance, achieving 94.45% Dice coefficient and 19.2% reduction in boundary localization error compared to state-of-the-art methods. Unlike prior work that applies these techniques independently, this work demonstrates that their adaptive, synergistic integration within a diffusion-based framework yields substantial improvements in boundary accuracy and topological correctness. 2026 The Author(s). -
Lalasa Quantum Computing Method: A Unique Quantum Convolutional Neural Network Architecture
A novel Lalasa Quantum Computing Method architecture is presented in this paper for classification of medical images, aiming to enable efficient early detection of cancer. The proposed framework integrates a custom preprocessing pipeline that removes specular reflections using a SWIN Transformer and segments regions of interest via an Enhanced Gaussian Mixture Model. The quantum classification module employs amplitude encoding to map classical image data into quantum states, enabling structured feature extraction through a sequence of quantum convolutional layers with trainable variational circuits. The model was implemented using IBM Qiskit and trained on the publicly available Intel & Mobile ODT Cervical Cancer Screening dataset. Experimental results show a high overall classification accuracy of 98.58%, with moderate performance on class-specific F1-scores, recall, and precision. These results demonstrate the feasibility and effectiveness of quantum-classical hybrid models for medical image analysis, particularly in high-dimensional, low-sample scenarios. The study sets stage for future advancements in quantum machine learning applications in healthcare, with potential extensions involving real quantum hardware deployment and multiclass classification improvements. 2025 IEEE. -
EGMM: removal of specular reflection with cervical region segmentation using enhanced Gaussian mixture model in cervix images
Colposcopy is a crucial imaging technique for finding cervical abnormalities. Colposcopic image evaluation, particularly the accurate delineation of the cervix region, has considerable medical significance.Before segmenting the cervical region, specular reflection removal is an efficient one. Because, cervical cancer can be found using a visual check with acetic acid, which turns precancerous and cancerous areas whiteand these could be viewed as signs of abnormalities. Similarly, bright white regions known as specular reflections obstruct the identification of aceto-whiteareas and should therefore be removed. So, in this paper, specular reflection removal with segmentingthe cervix region ina colposcopy image is proposed. The proposed approach consists of two main stages, namely, pre-processing and segmentation. In the pre-processing stage, specular reflections are detected and removed using a swin transformer. After that, cervical regions are segmented using an enhanced Gaussian mixture model (EGMM). For better segmentation accuracy, the best parameters of GMM are chosen via the adaptive Mexican Axolotl Optimization (AMAO) algorithm. The performance of the proposed approach is analyzed based on accuracy, sensitivity, specificity, Jaccard index, and dice coefficient, and the efficiency of the suggested strategy is compared with various methods. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024. -
AMAA-GMM: adaptive Mexican axolotl algorithm based enhanced Gaussian mixture model to segment the cervigram images
Colposcopy is a crucial imaging technique for finding cervical abnormalities. Colposcopic image evaluation, particularly the accurate delineation of the cervix region, has considerable medical significance. Before segmenting the cervical region, specular reflection removal is an efficient approach. Cervical cancer can be found using a visual check with acetic acid that turns precancerous and cancerous areas white and these could be viewed as signs of abnormalities. Similarly, bright white regions known as specular reflections obstruct the identification of aceto-white areas and should therefore be removed. So, in this paper, specular reflection removal with segmenting the cervix region in a colposcopy image is proposed. The proposed approach consists of two main stages, namely, pre-processing and segmentation. In the pre-processing stage, specular reflections are detected and removed using a swin transformer. After that, cervical regions are segmented using an enhanced Gaussian mixture model (EGMM). For better segmentation accuracy, the best parameters of GMM are chosen via the adaptive Mexican axolotl optimisation (AMAO) algorithm. The performance of the proposed approach is analysed based on accuracy, sensitivity, specificity, Jaccard index, and dice coefficient, and the efficiency of the suggested strategy is compared with various methods. Copyright 2026 Inderscience Enterprises Ltd. -
A Novel Cross-Validation Fusion Model Combining Vision Transformer and DenseNet161 for Enhanced Cervical Lesion Classification
Cervical cancer is fourth most common cancer in women across the world with highest impact in low- and middle-income countries. World Health Organization sent out a call for all UN nations to work toward the elimination of cervical cancer. Deep learning and artificial intelligence have been the go-to solutions for medical image analysis for diagnosis and prognosis. This paper aims to classify lesions in a colposcope captured cervix image with help of artificial intelligence models. To further advance automated cervical lesion classification, the study proposes a novel hybrid model that combines the complementary strengths of a vision transformer and DenseNet architecture. The paper also addresses ongoing challenges, such as interference from specular reflection areas and the difficulty in distinguishing between different lesion grades due to subtle visual differences. The proposed cross-validation decision fusion strategy aims to improve the reliability and robustness of the classification process. The results of the study affirm that deep learning and fusion technologies will steer the future direction of research in medical image analysis. DenseNet model has performed with an accuracy score of 0.695, sensitivity of 0.912, specificity of 0.979 and F1 score of 0.9100. These metrics are significantly improved versions of state of the art used in this study for comparative analysis. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Comparative Performance Analysis of Segmentation Methods in Cervigram Images
One of the most common cancers of the lower female reproductive tract is cervical cancer and it is a major contributor of mortality in developing nations. Screening tests include image analysis of pap smear and colposcope pictures. In image analysis, machine learning techniques can be employed to analyze and interpret images of the cervix through segmentation and extraction of characteristics for the classification of cervix images. K-means algorithm and Gaussian mixture model are popular segmentation algorithms used in cervix region-of-interest extraction. In the context of deep network learning, segmentation means the use of deep convolution networks to accurately identify different objects or regions in an image. R-CNN and Deeplab architectures are among the most frequently employed models in deep learning for automated cervix image processing. In this paper, we have systematically reviewed machine and deep learning models popularly employed in cervical cancer identification through colposcope images. Four carefully chosen models were deployed, and their performance was comparatively analyzed. This research can be a foundation for scientists looking to develop new models for the classification and segmentation of cervical cancer. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Enhancing Mobility: A Smart Cane with Integrated Navigation System and Voice-Assisted Guidance for the Visually Impaired
Blindness is a condition which affects many people, and for the affected people, quality of life can take a big hit. Most blind people already use walking sticks to feel the terrain in front of them as they move around and navigate using touch and sound. However, they cannot judge distances to objects until the cane actually hits the object. In some cases, the contact with the cane may damage the object. Hence, it may be better to have some early warning system so that there is less likelihood of causing damage. This paper presents the design and development of a 'Smart Cane' aimed at enhancing mobility and safety for visually impaired individuals. The cane incorporates ultrasonic sensors to detect objects in the user's surroundings. When an object is detected within a specified distance range, the cane provides haptic feedback through a bidirectional vibration motor, alerting the user to its presence. The microcontroller-based system processes data from both sensors and efficiently manages power consumption to ensure extended battery life. The device's design includes user-friendly controls and an ergonomic enclosure to offer ease of use and protection for the electronic components. Further, there is built-in navigation via online Map API. With the convenience of navigating oneself without external assistance, the 'Smart Cane' demonstrates great potential to improve the independence and confidence of visually impaired individuals in navigating their environments safely. 2024 IEEE. -
-
Comparative Analysis of GANs and Diffusion Models for Hyperspectral Image Classification
Hyperspectral imaging, which is obtained across numerous spectral bands, presents difficulties in classification due to its high dimensionality and intricate nature. This study provides a comparison of Generative Adversarial Networks (GANs) and Diffusion models regarding the classification of the Indian Pines, Pavia University, and Salinas Datasets, utilizing Multi-Layer Perceptron and Random Forest classifiers. The findings indicate the GANs combined with Random Forest outperform Diffusion models, attaining accuracies of 88%, 96% and 95% respectively. This approach may not outperform the top models, such as HTD-2D-3D-PCNN, but is simpler in structure and more computational efficient. Key recommendations would be real-time processing, edge device optimization, and applications customized to agriculture and urban planning. The Author(s), under exclusive license to Springer Nature Switzerland AG 2026. -
Predicting the financial behavior of Indian salaried-class individuals
COVID-19 has caused not only unprecedented health crises but also economic crises among individuals across the world. White-collar (salaried-class) employees with a fixed salary face financial insecurity due to job loss, pay cuts and uncertainty in retaining a job. This study examines the financial behavior of Indian white-collar salariedclass investors to their cognitive biases. In addition, the mediating effect of financial self-efficacy on cognitive biases and financial behavior is examined. Respondents were given structured questionnaires (google forms) through emails and WhatsApp for data collection. SPSS and R-PLS are used to analyze the data. Conservatism (r = -.603, p < 0.05) and herding bias (r = -.703, p < 0.05) have a significant negative correlation with financial behavior. Financial self-efficacy has a significant positive correlation (r =.621. p < 0.050). Conservatism and herding predicted 60.5% and 62.2% of the variance, respectively. The direct and indirect paths between conservatism bias, financial self-efficacy, and financial behavior are significant. The paths between herding, financial self-efficacy and financial behavior are also significant. Ankita Mulasi, Jain Mathew, Kavitha Desai, 2022. -
Impact of Behavioural Biases, Emotional Intelligence and Financial Literacy on Financial Behaviour
Behavioural finance is the integration of finance and psychology. Finance pertains to the administration of finances by an individual, while psychology involves the examination of the mind and human conduct. Behavioural finance provides a reasonable explanation for the abnormalities observed in financial markets and the irrational decisions made by investors. The foundation of behavioural finance lies in the irrational choices made by individuals, which elucidates the reasons for investors frequently encountering different biases, emotional filters, and a lack of knowledge when making financial decisions. The primary challenges stem from three critical deficiencies: the investors' inability to express their cognitive and emotional biases, insufficient financial knowledge, and an inability to regulate emotions during financial decision-making. This study investigates the influence of investors' financial literacy, emotional intelligence, and behavioural bias on their financial conduct. Studies have shown that behavioural biases negatively impact financial behaviour. Financial literacy and emotional intelligence have been recognised as critical elements that significantly impact the financial behaviour of salaried class investors. -
Leadership strategies for change and innovation
Organizational innovation and change development depends heavily on leadership activities. This chapter investigates how leaders should integrate charismatic and authentic methods with transformational concepts to create the best possible leadership practice for long-term organizational success and innovation promotion. The chapter also demonstrates that leadership success will emerge from blending different leadership styles since any one approach alone is inadequate for solving current organizational challenges. Strong leaders who learn to unite multiple leadership approaches gain enhanced abilities to encourage their personnel while promoting innovative initiatives. 2025, IGI Global Scientific Publishing. All rights reserved. -
JP-DAP: An Intelligent Data Analytics Platform for Metro Rail Transport Systems
This paper deals with an intelligent data analytics platform-Jaison-Paul Data Analytics Platform (JP-DAP)-for metro rail transport systems. JP-DAP is intended to ensure smooth functioning, improved customer experience, ridership forecasting, and efficient administration of metro rail transportation systems by integrating and analysing its many data sources. It consists of a middleware which is built on the top of a Hadoop Distributed File System (HDFS) and Spark framework, along with a set of open-source software tools like Apache Hive, Pandas, Google TensorFlow and Spark ML-lib for real-time and legacy data processing. The benchmarking of JP-DAP was conducted using TestDFSIO and have found that it performs well according to industry standards. The specific use case for this project is Kochi Metro Rail Limited (KMRL). The analysis of Automated Fare Collection data from KMRL on JP-DAP framework have produced descriptive statistics visualisation of inflow and outflow analysis, travel patterns during weekdays and weekends, origin-destination matrix, etc.. Moreover JP-DAP framework is capable of producing short term passenger flow predictions using SVR machine learning algorithm with linear, radial basis function and polynomial kernels. Our experiments have shown that SVR linear kernel gives the most accurate results with the least errors in predicting the next day's passenger count using the previous five weekdays data. The station usage (one-to-all) prediction using Long Short-Term Memory (LSTM) is also integrated to this framework. The visualisation as well as analytical outcomes of JP-DAP framework have also been made available to the external world using a rich set of REST APIs and are projected on to a web-dashboard. 2000-2011 IEEE. -
Comparative Analysis of Neural Network Models for Indian Sign Language Hand Gesture Recognition
The recognition of sign language is a crucial element in filling communication gaps that exist in the population. As inclusive communication technologies become more popular, there has been a significant push to develop trustworthy systems for translating sign language into written or visual form. The use of hand gestures and body movements is a fundamental aspect of sign languages, which are commonly used by those who are deaf. The lack of proficiency in sign language makes communication difficult for most people. A project was undertaken to convert Indian Sign Language (ISL) into spoken language through research. The paper presents a comparison of various neural network models. Using OpenCVgenerated real-time images and MediaPipe, it is possible to identify hand movements and collect ISL gesture data in realtime. In the study, it was demonstrated that ResNet50 is 92 per cent accurate in real-time recognition when compared to other models. This work aims to promote inclusivity and communication skills among people who may not have the ability to hear or speak fluently. Adding face recognition to future work may improve accuracy and enable continuous sign language recognition, providing more dynamic and real-time translation capabilities. 2025 IEEE. -
Restrained geodetic domination of edge subdivision graph
For a connected graph G = (V,E), a set S subset of V (G) is said to be a geodetic set if all vertices in G should lie in some u-v geodesic for some u,v S. The minimum cardinality of the geodetic set is the geodetic number. In this paper, the authors discussed the geodetic number, geodetic domination number, and the restrained geodetic domination of the edge subdivision graph. 2022 World Scientific Publishing Company.