Category Archives: Uncategorized

Authors: Mrs.KanakaTulasi P.Reddi, Sai Varshitha Kuppili, Gabu Ganesh Sasikanth, Adapa Sai Teja Venkata Vinay, Medaboyina Karthik, Trivinesh Gundra

Abstract: Autonomous marine vessels rely heavily on artificial intelligence systems for accurate sea state estimation, which plays a crucial role in navigation, stability control, and operational safety. However, AI-based models are vulnerable to adversarial attacks, where small and carefully crafted perturbations in input data can significantly degrade model performance. Such attacks may compromise safety and reliability, especially in critical maritime environments.This project proposes a novel robustness-enhancing adversarial defence approach to improve the reliability of AI-powered sea state estimation systems. The framework focuses on strengthening deep learning models against adversarial perturbations while maintaining high estimation accuracy. The system integrates adversarial training and defensive mechanisms to enhance model stability under uncertain and hostile conditions. Experimental evaluation demonstrates that the proposed defence strategy significantly improves robustness without sacrificing predictive performance. The results confirm that the enhanced model maintains reliable sea state estimation even in the presence of adversarial disturbances.The proposed approach contributes to improving the safety, security, and reliability of autonomous marine navigation systems.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.152

Published by: vikaspatanker

Authors: Mrs.M.Uma Devi, Togaru Reshma Sri, Katikidala Satya Ratna Naveen, Gubbala Leela Madhavi, Kalvakolanu Venkata Pavan Chaitanya, Velduti Srivenkata Surya Sai Kumar

Abstract: The rapid advancement of generative artificial intelligence has made it increasingly difficult to distinguish between real images and AI-generated synthetic images. Modern diffusion models can produce highly realistic visuals that closely resemble authentic photographs, raising serious concerns about misinformation, digital fraud, and media manipulation. As synthetic image generation becomes more accessible, reliable detection mechanisms are essential to maintain digital trust and security.This project presents an image classification framework for identifying AI-generated synthetic images using deep learning techniques. A balanced dataset is constructed by combining real images from the CIFAR-10 dataset with synthetic images generated using Stable Diffusion. A Convolutional Neural Network (CNN) model is trained to perform binary classification, distinguishing between real and fake images. In addition to classification, Explainable Artificial Intelligence (XAI) techniques such as Grad-CAM are applied to interpret model decisions and visualize the regions that influence predictions.Experimental results demonstrate that the proposed model achieves high accuracy in detecting synthetic images while maintaining reliable generalization performance. The explainability component further enhances transparency by revealing distinctive patterns and artifacts present in AI-generated images. The proposed system contributes to improving digital image forensics and strengthening defences against AI-driven visual misinformation.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.151

Published by: vikaspatanker

Authors: Dr. Y V Ram Kumar, Keerthi Teja Sri, Patamsetti Yaga Sri Satya Prasanna, Chintapalli Sashank, Theripogu John, Surapureddy Venkata Sai Praneeth

Abstract: Accurately estimating traffic states is a crucial aspect of transportation engineering, enabling effective traffic control and operations. In recent years, Physics-Regulated Deep Learning (PRDL) has gained significant attention due to its ability to achieve higher accuracy while requiring less training data compared to conventional deep learning (DL) approaches. However, a key challenge of PRDL is the lengthy training time required for closely related but distinct tasks.To address this limitation, this paper introduces a hybrid physics-regulated deep transfer learning approach that leverages the strengths of transfer learning, PRDL, and DL to enhance estimation accuracy and reduce computational costs, particularly in scenarios with limited observation data. The proposed framework includes two transfer learning variants designed to extract and transfer essential features from pre-trained models to new but similar traffic environments. This hybrid approach integrates deep learning training, minimizing computational overhead by eliminating physics-based loss calculations during training.Simulation results demonstrate that, compared to traditional PRDL methods, the proposed transfer learning approaches improve estimation accuracy by over 12% on average while reducing training time by more than 50% on average. These findings highlight the potential of hybrid transfer learning techniques in accelerating the adoption of PRDL for traffic state estimation, making it a valuable tool for transportation systems with limited computational resources.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.150

Published by: vikaspatanker

Authors: Dr. K. Mounika, Rachakonda Shashikanth, Palivela Prem Chandu, Virothula Sasi Kumar, Balusu Eswar, Kandikonda Pranay

Abstract: The accurate forecasting of Daily Global Solar Radiation (DGSR) is a vital tool in renewable energy planning, climate research, and environmental monitoring. This paper presents a proposal for utilizing machine learning to enhance the estimation of DGSR by using satellite image data. This method utilizes reflectance values obtained from Metaset Second Generation (MSG) satellite images across multiple spectral channels and relies on ground-based meteorological parameters rather than traditional models. Artificial Neural Network (ANN) and Support Vector Machine (SVM) are two supervised machine learning regression models that are utilized for forecasting solar radiation. How do they compare and contrast? The Gharda radiometric station in Algeria has collected measurable solar radiation data for four years (2014-2017) using inputs from satellite imagery, which are then combined to create the dataset. To evaluate these models, statistical performance metrics such as Root Mean Square Error (RMSE), Normalized RMSEA (NRMSE) and MAE (Made Absolute Percentage Error), MBE (Merck-McGregor), and correlation coefficient (R) are utilized. Prediction accuracy is significantly influenced by the number and combination of satellite input parameters, as demonstrated by experimental data. Compared to the SVM, the ANN model had a better RMSE of 21221. The NRMSE, MAPE, and MBE have all been reported with 3.46%, 2.85%, 7.26, etc, respectively. Wh/m2. A 0.99 correlation coefficient is associated with a Wh/m2 value.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.149

Published by: vikaspatanker

Authors: Mr.V.Prem Kumar, Manyam Teja Siva Ganesh Goud, Kothapalli Vennela Sri Sai Bhargavi, Yeluri V N S S P Teja, Vedagiri Yuva Sai Suresh, Dara Teja

Abstract: Traffic violations have become a major cause of road accidents and fatalities in many countries, particularly in densely populated urban areas. Common violations such as red-light jumping, triple riding on two-wheelers, and reckless driving significantly increase the risk of road accidents. Traditional traffic monitoring systems rely heavily on manual observation by traffic police or limited sensor-based systems, which are inefficient, time-consuming, and prone to human errors. To address these challenges, intelligent traffic monitoring solutions based on computer vision and deep learning have gained significant attention. This paper proposes a deep learning-based automated traffic violation detection system using the YOLOv7 object detection model. The proposed system processes video streams obtained from roadside surveillance cameras and analyses them frame-by-frame to detect different traffic violations. The YOLOv7 model is employed to identify vehicles and generate bounding boxes around detected objects. A predefined threshold line is used to determine whether a vehicle crosses the traffic signal during a red light, thereby identifying signal violations. Additionally, the system detects over boarding or triple riding on two-wheelers by analysing the number of riders detected within a single vehicle bounding box. The system uses publicly available datasets such as the MS COCO dataset for vehicle detection and a custom annotated dataset for over boarding detection. The model is trained and evaluated using performance metrics including precision, recall, F-measure, and mean Average Precision (mAP). Experimental results demonstrate that the proposed model effectively detects multiple traffic violations with high accuracy while maintaining efficient real-time performance. The proposed approach provides a cost-effective, automated, and scalable traffic monitoring solution that can assist traffic authorities in improving road safety and reducing the workload associated with manual monitoring systems. The system can be integrated with existing smart city surveillance infrastructures to enhance intelligent transportation management and law enforcement.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.148

Published by: vikaspatanker

Authors: Mrs.K.Ganga Devi Bhavani, Bonam Geetha Chitti Jyothi, Siravapu Santhi Kumari, Karri Manikanta Sai, Alluri Sri Akshay Satya Srinivas, Thella Aditya

Abstract: The spread of fake news has become a significant concern in today’s society, as misleading information can easily damage reputations and lives. To address this issue, researchers have developed fake news detection systems using machine learning techniques. The identification of fake news is rapidly gaining traction and is increasingly being adopted by various industries, either for their own use or to offer as a service to others. Machine learning (ML) and deep learning (DL) are two prominent approaches employed to determine the authenticity of news. There are various methods available for detecting false news through both ML and DL techniques. This paper presents a comprehensive analysis of fake news detection using machine learning approaches. Upon thorough examination, it was found that several ML and DL algorithms have been applied in this domain, with the Support Vector Machine (SVM) being the most commonly used ML method, and Long Short-Term Memory (LSTM) being the most widely applied DL technique.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.147

Published by: vikaspatanker

Authors: Ms.G.Naga Rani, Mangipudi V N S Sekhar Sarma, Khandavalli V V Lakshmi Srirama Karthik, Malla Karthik, Pabbineedi Vanshika, Ventru Hemanth Kumar

Abstract: The banking sector plays a vital role in the global financial system by providing loans to individuals and businesses for various purposes. While loans generate significant revenue through interest, there is always a risk that borrowers may fail to repay the loan, resulting in financial losses for lending institutions. Therefore, accurately predicting the risk level associated with a loan application is an important task for banks and financial organizations. Traditional loan approval processes rely heavily on manual analysis of customer information, which can be time-consuming and prone to human bias. With the advancement of machine learning techniques, automated systems can now analyse large amounts of financial data to support more efficient and accurate loan approval decisions. This study proposes a machine learning-based loan risk prediction system that analyses customer personal and financial attributes to determine the likelihood of loan default. The dataset used for this study contains multiple features commonly included in loan applications, such as credit history, checking account status, loan amount, employment status, and age of the applicant. Data preprocessing techniques including outlier removal, categorical encoding, and feature scaling are applied to prepare the dataset for model training. Several machine learning algorithms are implemented and compared, including Decision Tree, Random Forest, Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), Naive Bayes, and a Stacking Ensemble model. The models are evaluated using performance metrics such as accuracy, precision, recall, F1-score, and confusion matrix analysis. Experimental results demonstrate that ensemble-based approaches provide improved predictive performance compared to individual machine learning models. The proposed system can assist financial institutions in making faster and more reliable loan approval decisions by identifying high-risk applicants before granting loans. By leveraging machine learning techniques, the system enhances the efficiency of credit risk assessment and supports more effective financial decision-making in the banking industry.

DOI: http://doi.org/10.61137/ijsret.vol.12.issue2.146

Published by: vikaspatanker

Authors: Poosarla Durga Bhavani, Shaik Roshan Jameer, Vasamsetti Monika Durga Satya Vani, Shaik Ubaid Ahamad, Purushottapatnapu Ravi Sai Krishna, Mr. A. V. Sudhakar Rao

Abstract: Finding a rental or a tenant is often a difficult task. The market is messy, with listings spread across many sites, often outdated, and communication is slow. "Rent Ease" is an all-in-one digital platform built to fix these problems, providing a single, reliable hub that makes renting simpler for everyone. This platform integrates modern technologies such as React.js for frontend development, FastAPI for backend services, and MongoDB for data storage. Key features include real-time chat communication between users and owners, Google Maps integration for accurate location tracking, and an AI-powered module that automatically generates property descriptions to enhance listing quality. Additionally, the system provides filtering options, detailed property views, and a rating and review mechanism to ensure transparency and better decision- making. This system improves user experience, reduces communication gaps, and provides reliable property information, making it a comprehensive solution for both property owners and tenants. Our main goal is to modernize the rental experience. By leveraging technology to connect owners and tenants directly, Rent Ease makes the process faster, more transparent, and significantly less of a hassle, creating.

DOI: https://doi.org/10.5281/zenodo.19451659

 

Published by: vikaspatanker

Authors: Piyush Pawar, Ayush Jagdale, Yuvraj More, Gunesh Padmukhe, Prof. Meshram A.G

Abstract: The Gesture Vocalizer is a smart assistive communication system developed to help speech- impaired and physically challenged individuals convey messages using hand gestures. The system employs gesture-detection sensors such as flex sensors or accelerometers to recognize predefined hand movements. These gestures are processed by a microcontroller, which converts them into corresponding voice outputs through a speaker or mobile application. The device enables real-time communication without the need for verbal speech, making it highly useful in daily interactions, hospitals, and emergency situations. Users can customize gesture-to- message mappings, improving flexibility and usability. By combining sensor technology, embedded systems, and voice output, the Gesture Vocalizer enhances independence, accessibility, and social interaction for differently-abled individuals.

DOI:

 

 

Published by: vikaspatanker

Authors: Ishita Shinde, Tanushri Jadhav, Apurva Ransing, Pritesh Patil, Dr. Mrunal Pathak

Abstract: The SkillBridge app aims to link professionals from a variety of industries with people who wish to acquire practical skills. Traditional learning approaches occasionally fall short of offering real-time mentoring and hands-on experience in today's quickly changing digital world. SkillBridge fills the need of developing a platform where students can find mentors, access skill-based resources and work together on learning opportunities. The software encourages knowledge sharing, community-driven learning, and skill development across a range of professions. Through the use of technology, SkillBridge assists professionals, students, and lifelong learners in increasing the effectiveness, accessibility, and interactivity of skill learning.

DOI: https://doi.org/10.5281/zenodo.19449850

Published by: vikaspatanker