Daily Archives: July 24, 2025

Authors: Himanshu Shahoo, GautamYadav, ChinmayeeTripathy, Padmaja Panda

Abstract: Flower identification is a crucial aspect of plant classification and ecological research, playing a significant role in understanding biodiversity and ecosystem dynamics. This research paper presents a new approach to flower identification using advanced deep learning techniques. The proposed system used folding networks (CNNs) to automatically extract hierarchical features from high-resolution images of flowers, allowing for more accurate and efficient classification. The procedure is implemented as a multi-stage process, beginning with data preprocessing to enhance image quality and remove noise. Using another data record, educated CNN models such as modified reset 50, VGG16, or Google are then fine-tuned with commented flower images. Furthermore, transfer learning is used to properly use knowledge from large data records and improve the ability of models to generalize different types of flowers.. In the end, our approach achieved an accuracy of 82.04% using VGG16, the highest compared to other algorithms.

DOI: http://doi.org/10.5281/zenodo.16610684

Published by: vikaspatanker

Authors: Yajat Singh, Ms. Gurpreet Kaur, Barun Singh Bisht, Pushpam Kumar

Abstract: Underwater Sensor Networks USNs play a critical role in environmental monitoring marine exploration and defense applications. However traditional cloudbased data processing introduces significant latency and energy consumption making realtime decisionmaking challenging in extreme underwater environments. This paper proposes a novel edge computing architecture tailored for USNs enabling localized realtime data processing and anomaly detection. The architecture integrates a CNNLSTM deep learning model optimized for lowpower edge devices significantly reducing the need for cloudbased processing. Our experimental evaluation demonstrates a 39 reduction in latency and a 36 improvement in energy efficiency compared to cloudbased solutions. Additionally we present performance benchmarks showing a higher packet delivery ratio and improved data throughput. The proposed approach enhances the autonomy and efficiency of underwater sensor networks making it a viable solution for realtime applications in extreme environments.

DOI: https://doi.org/10.5281/zenodo.16610646

 

Published by: vikaspatanker

Authors: Dr Vinit Kumar Sharma,, Anjali Goyal, 3Kushagra Sharma, Kushagra Sharma

Abstract: In this paper, we have discussed application of various methods ([9],[10]) for solving the problem of a game as Dominance method, Graphical method, Algebraic method, Simplex method etc. Each method has its limitations and benefits, which depends upon the nature of problem. Students may learn about the uses of various methods by study this paper.

Published by: vikaspatanker

Authors: Er. Rajdeep Saharawat,, Muskan,, Dr. Vinit Kumar Sharma,, Ms. Meenal Maan

Abstract: Proteins must fold into specific three-dimensional structures to function correctly. Errors in protein folding—misfolding—can lead to aggregation and are associated with several degenerative diseases, including Alzheimer’s, Parkinson’s, and Huntington’s. This review explores the molecular mechanisms of protein folding and misfolding, the cellular quality control systems managing these processes, and the pathogenesis of misfolding-related disorders. We also discuss therapeutic approaches aimed at correcting misfolding or enhancing proteostasis

Published by: vikaspatanker

Authors: Usha Dhankar, Nikeeta, Sompriya N Tiwary, Suhani Singh, Pooja Sharma, AS Susanna Grace

Abstract: Solar photovoltaic (PV) panels were a broadly implemented renewable energy source but their efficiency was substantially influenced by dust accumulation which hindered sunlight absorption and reduced power output Regular cleaning and monitoring were essential to sustain their performance Traditional cleaning mechanisms such as manual or semiautomatic cleaning were often inefficient labor-intensive and costly which demanded the development of automated solutions This research introduced an IoTbased cleaning and monitoring system designed to enhance the efficiency of solar PV panels The system combined realtime data acquisition through IoT sensors to detect dust accumulation and environmental conditions activating an automated cleaning mechanism when necessary Additionally machine learning algorithms analyzed historical data to optimize cleaning schedules maintaining minimal energy loss and improved reliability A review of prevalent dust removal techniques such as passive coatings electrostatic cleaning and robotic solutions revealed that many methods were either highmaintenance or not costeffective for largescale deployment IoTbased solutions when integrated with predictive analytics provided a potential substitute by enabling realtime monitoring and analytical decisionmaking for panel maintenance The outlined methodology enhanced energy output while reducing operational costs and minimizing manual intervention making solar energy systems more efficient and sustainable This innovation contributed to the prolonged effectiveness of solar power by addressing one of its key operational challenges thereby fostering a cleaner and more reliable renewable energy future.

DOI: https://doi.org/10.5281/zenodo.16410228

 

Published by: vikaspatanker

Authors: Kushagra Sharma, Dr Vinit Kumar Sharma, Anjali Goyal

Abstract: In this paper, we have discussed application of various methods ([9],[10]) for solving the problem of a game as Dominance method, Graphical method, Algebraic method, Simplex method etc. Each method has its limitations and benefits, which depends upon the nature of problem. Students may learn about the uses of various methods by study this paper.

Published by: vikaspatanker

Authors: Senthil Kumar T, Vedant Kolhe, Rithvee Chandak

Abstract: Recently mental health has been regarded as an important issue with stress being one of the factors behind many health conditions. Prompt detection of mental stress is critical in preventing chronic conditions. Artificial intelligence has been helping to fight against chronic stress and tension. This study provides a review of the current understanding of stress and artificial intelligence as well as the approaches for overcoming it with AI algorithms. Some of the approaches studied in development include LSTM networks self-organizing maps and natural language processing applied to datasets. The comparative analysis of these methods enables us to determine the most successful approaches their limitations and ways in which they can be improved.

DOI: https://doi.org/10.5281/zenodo.16410131

 

Published by: vikaspatanker

Authors: Vishal Kumar Pandey, Vishal Jaiswal, Vishal Yadav, Shilpee Patil

Abstract: Classroom attendance tracking was a fundamental task in educational institutions, traditionally managed through manual roll calls or sign-in sheets. These methods were time-consuming, error-prone, and susceptible to manipulation. With advancements in computer vision and embedded systems, there was an opportunity to automate this process. In this research paper, a novel approach to classroom attendance management was presented, utilizing OpenCV and face recognition technologies, implemented on the ESP32-CAM microcontroller. The proposed system was designed to automatically identify and record student attendance, offering enhanced accuracy and efficiency. Comparative results demonstrated that the face recognition-based approach significantly outperformed traditional manual methods and other automated systems in terms of accuracy and processing speed. The system's architecture, implementation, and evaluation were outlined, showcasing its potential to transform attendance tracking in educational settings.

DOI: https://doi.org/10.5281/zenodo.16410064

 

Published by: vikaspatanker

Authors: Ms. Aneesha Shokeen, Mr. Yash Mittal

Abstract: In emergency situations, quick response and hands-free communication are critical for safety. This research introduces an AI-driven gesture recognition system designed to activate an SOS alert using simple hand gestures. The system employs MediaPipe for precise hand tracking, an SVM classifier for real-time gesture recognition, and an integrated SMS alert system that includes GPS location tracking. The proposed approach ensures accessibility for individuals in distress, particularly those with disabilities or in high-risk environments where traditional emergency triggers may be impractical. The experimental results show a 95.19% accuracy in gesture classification, demonstrating the system’s effectiveness in real-world scenarios. Future work aims to incorporate deep learning-based recognition models and deploy the system on wearable and mobile platforms to enhance usability and responsiveness

DOI: https://doi.org/10.5281/zenodo.16410503

 

Published by: vikaspatanker

Authors: Anant Samrat, Amar Deep Gupta, Shubham Dadwal, Adarsh Samrat,, Mayank, Priya Kumari

Abstract: Crime is a significant challenge in modern society, necessitating effective prevention strategies. Machine learning (ML) offers promising solutions for crime analysis and prediction. This study explores algorithms like Naive Bayes, SVM, Linear Regression, Decision Trees, Bagging, Stacking, and Random Forest for accurate crime prediction. The proposed Naive Bayes-based model achieved 99.9% classification accuracy on test data, outperforming previous models. By integrating empirical data and criminological insights, this approach effectively forecasts crimes, reducing crime and deterring criminal activities.

DOI: https://doi.org/10.5281/zenodo.16409884

 

Published by: vikaspatanker