Category Archives: Uncategorized

Authors: Rohit.N, Dr.R. Kannadasan

Abstract: In human anatomy, the liver has a special feature called regeneration, and this feature helps the liver to grow back even after a large part of its organ is removed. Like regeneration, it helps maintain bile salts, protein synthesis, and detoxification. Irregular eating habits, sleep, and alcohol consumption increase liver function and cause various diseases such as fatty liver and other liver problems. Hepatocellular carcinoma is one of the liver diseases, and it is caused by the abnormal growth of cells in the liver. In such cases, liver regeneration is possible if the disease is identified at an early stage. To support this early identification, several research works have been carried out using both artificial intelligence and deep learning techniques. Therefore, this paper proposed an automated approach to identify liver cancer at an early stage through a segmentation and classification using deep learning techniques. The early-stage identification becomes possible with the dual stage of segmentation using U-NET and XAI. The U-NET helps to segment the image through its various texture properties of the image. Then, the XAI is used to analyze the individual regions of the image. This special feature of this approach is that it uses descriptive AI for classification, and this helps in identifying critical regions through heat maps and saliency maps. This technique was tested on a computed tomography dataset of liver images and its performance was evaluated in terms of precision, accuracy, recall and F1-score.

DOI: https://doi.org/10.5281/zenodo.17696117

Published by: vikaspatanker

Authors: V.Gomathi priya

Abstract: Disaster management has transformed from a reactive emergency-centred practice to a proactive, technology-driven discipline focused on minimizing the impact of both natural and human-induced hazards. This research highlights the rising frequency and severity of disasters such as earthquakes, floods, and cyclones, emphasizing the demand for integrated, systematic, and multi-sectoral strategies. The study examines how contemporary frameworks now prioritize preparedness through scenario planning, prevention, and capacity-building, supplemented by robust early-warning systems and risk reduction measures. The role of advanced technology—including decision support systems powered by artificial intelligence, geographic information systems, and, more recently, quantum computing—is analysed for its effectiveness in rapid data processing, predictive modelling, and optimized resource allocation. The research also reviews key disaster events from the past five years to illustrate ongoing challenges and responses, demonstrating the need for resilient infrastructure and coordinated recovery operations. Findings suggest that modern disaster management not only saves lives and reduces economic loss during crises but also promotes societal resilience and sustainable development through continuous innovation, capacity enhancement, and collaborative governance. This evolution ensures that disaster management remains a critical element for safeguarding communities in an era of mounting environmental and technological risk.

Published by: vikaspatanker

Authors: Priyanka K, Sasiprabha V, Shobaa P, Shreelinganaathan M, Srimukhi TG

Abstract: Low‑voltage distribution lines are widely used in residential and semi‑urban areas, where accidental conductor breakage or unintended grounding can pose severe safety hazards, result in prolonged power outages, and damage equipment. Current detection methods rely heavily on manual handling or high‑cost protection systems, resulting in delayed responses and increased operational risks. To address these gaps, this work proposes a cost‑effective Smart LV Fault Detection System capable of identifying conductor breakage and earth faults in real time. The system integrates voltage and current sensing units with an ESP32 controller that continuously analyses line behaviour, recognises abnormalities, and triggers immediate alerts. Visual and audio warnings are activated on-spot, while remote fault notifications are delivered via GSM‑based communication. A solar‑powered backup ensures uninterrupted operation during power failures. Experimental tests show more than 95% detection accuracy with a response time under one second. The proposed solution is lightweight, scalable, and suitable for both rural and urban power networks, providing improved safety, reduced downtime, and faster maintenance response.

Published by: vikaspatanker

Authors: Mr. S. Kaushik Raj, Mrs. B. Shyamala Devi

Abstract: Diabetic retinopathy (DR) is one of the major causes of vision loss worldwide, making early and reliable detection extremely important. This work presents an advanced transformer-driven framework combined with a Multiple Instance Learning (MIL) strategy to classify DR using retinal fundus images. The transformer model effectively learns long-range relationships and contextual patterns, while the MIL approach analyzes image patches to highlight clinically significant areas. Together, this hybrid system delivers strong feature representation and stable classification performance, even with variations in image quality and resolution. Trained on a large and diverse dataset, the proposed model achieves higher sensitivity and specificity than many existing deep learning techniques. The system is designed to assist eye-care professionals by enabling accurate, timely assessments and providing a scalable solution suitable for extensive DR screening initiatives.

DOI: https://doi.org/10.5281/zenodo.17680366

Published by: vikaspatanker

Authors: Mr.S. Sathish, Ms. A. Sangeetha

Abstract: Object detection is a key area in computer vision with wide-ranging applications such as autonomous driving, surveillance, and augmented reality. YOLOv8, an advanced version of the YOLO series, stands out for its high accuracy and real-time performance. This study focuses on the analysis and implementation of YOLOv8 for real-time object detection, emphasizing its architecture that employs a deep neural network to perform a single forward pass for predicting bounding boxes and class probabilities simultaneously. The model’s main components—backbone network, detection layers, and anchor boxes—work together to achieve fast and efficient detection. Practical aspects such as model optimization, GPU acceleration, and post-processing are also explored to enhance speed and accuracy. Experiments conducted on benchmark datasets and real-world data demonstrate YOLOv8’s effectiveness, proving it to be a robust and adaptable solution for real-time object detection tasks. This research contributes to the advancement of computer vision and provides practical insights for deploying YOLOv8-based detection systems across multiple domains.

DOI: https://doi.org/10.5281/zenodo.17680246

Published by: vikaspatanker

Authors: Dr.P.Guhan, Mr.C.Barath

Abstract: Gesture-controlled computers and laptops have recently become increasingly popular, with Leap Motion technology leading this innovation. This technique enables users to control certain system functions simply by moving their hands in front of a camera. Compared to traditional slides or overhead projectors, computer-based presentations offer greater interactivity through audio, video, and programmable elements, though they can be more complex to use. As technology continues to evolve, finding new and affordable ways to interact with computers has become essential, especially since touchscreens are not feasible for all applications. To address this, a virtual mouse system based on object tracking and hand gestures is proposed as an alternative to physical mice and touch interfaces. The system employs computer vision techniques using Python and OpenCV, with a webcam detecting hand movements through HSV color segmentation. Users can wear colored caps or tapes on their fingers to move the cursor and perform actions like left-click, right-click, and double-click. The camera feed is processed in real time and displayed on the screen, allowing smooth, contactless interaction.

DOI: https://doi.org/10.5281/zenodo.17680123

Published by: vikaspatanker

Authors: Sabitha.K, Mohan Sundar.M, Santhosh. A. S, Sri Raghav. B. V, Supriya.A

Abstract: The Vehicle Accident Detection and Notification App is a mobile safety tool designed to identify vehicle accidents and automatically notify emergency services and family members. Utilizing the smartphone's accelerometer and gyroscope sensors, the app tracks sudden impacts, sharp turns, or rapid stops that suggest a potential collision. Upon detecting unusual movement, the GPS feature activates to pinpoint the exact location, ensuring that emergency responders can get to the scene quickly, even if the user is unconscious. The app automatically sends a comprehensive alert message that includes the user's information, timestamp, and precise coordinates to registered contacts. To minimize false alarms, it incorporates a brief confirmation timer that enables the user to cancel the alert. Additional functionalities such as a manual SOS button, cloud-based data storage, and automated calling improve reliability. Overall, the app provides an efficient, affordable way to enhance road safety and decrease fatalities.

Published by: vikaspatanker

Authors: Himanshu Kumar Rai, Harsh Goyal, Dr. Anshu Gupta

Abstract: This research proposes a Geo‑AI based smart attendance monitoring system that will integrate Artificial Intelligence (AI) with Global Positioning System (GPS) geofencing to provide accurate, contactless, and secure attendance tracking. The proposed system will address persistent challenges such as proxy attendance, location spoofing via mock‑GPS tools, and manual recording errors by combining AI‑based facial recognition (with liveness analysis) and geolocation verification. A mobile front end will enable frictionless check‑ins, while a cloud back end will ensure secure storage, auditability, and real‑time analytics. We detail the proposed architecture, algorithms, and implementation plan that will use a React Native application, Node.js/Express services, MongoDB storage, and Python micro‑services for inference and anomaly detection. If successfully implemented, in a controlled deployment across two office sites the system is expected to achieve 98% attendance‑marking accuracy, is projected to reduce administrative overhead by approximately 60%, and is anticipated to deliver alert latencies below five seconds. We will compare our approach with RFID, fingerprint biometrics, and GPS‑only mobile apps, and we will report an ablation analysis to quantify the benefits of liveness checks and geofence validation.’

DOI: http://doi.org/10.5281/zenodo.17679353

Published by: vikaspatanker

Authors: Manisha Rai, Vishal Rai

Abstract: Human emotions exhibit complex, nonlinear behaviors that traditional linear psychological models fail to capture adequately. This paper proposes a comprehensive framework for understanding emotional dynamics through the lens of chaos theory and nonlinear dynamical systems. We examine how concepts such as strange attractors, bifurcations, Lyapunov exponents, and phase space representations can illuminate the intricate patterns underlying affective processes. Empirical evidence from heart rate variability studies, electroencephalographic analyses, and longitudinal mood assessments supports the view that healthy emotional functioning corresponds to a bounded chaotic regime characterized by optimal complexity and adaptability. We further explore how deviations from this regime—manifesting as either excessive rigidity or instability—may underlie various mood disorders including depression, anxiety, and bipolar disorder. The paper presents mathematical formulations for emotional phase space dynamics, discusses computational approaches for reconstructing emotional attractor landscapes from empirical data, and outlines applications in clinical psychology, affective computing, and personalized mental health interventions. Understanding emotions as emergent properties of complex dynamical systems offers profound implications for diagnosis, treatment, and the development of emotionally intelligent technologies.

Published by: vikaspatanker

Authors: Mr. Mayank Arora, Nupur

Abstract: The rapid expansion of digital technologies has transformed the landscape of contemporary journalism, bringing data-driven reporting to the forefront of news production. In India, the rise of data journalism—an approach that integrates statistical analysis, visualization tools, and storytelling—has opened new possibilities for accuracy, depth, and transparency in media reporting. Yet, the adoption of data journalism remains uneven, complicated by structural issues within Indian newsrooms, limited technological expertise, and the pressures of fast-paced news cycles. This paper investigates how Indian news organizations understand, adopt, and implement data journalism practices. It explores the professional, infrastructural, and ethical challenges that constrain data-driven reporting, while also identifying opportunities created by digital literacy, open-data movements, and audience demand for evidence-based journalism. Through a review of scholarly literature, industry reports, and comparative perspectives, the study highlights how data journalism in India stands at a critical intersection of innovation and limitation. The paper argues that although data journalism has the potential to strengthen public discourse and democratic accountability, its growth depends on sustained investment in training, technological resources, and editorial vision. Ultimately, the study positions data journalism as an evolving journalistic paradigm that can contribute significantly to India’s media ecosystem if supported by a culture of transparency, collaboration, and professional development.

Published by: vikaspatanker