Daily Archives: February 10, 2026

Authors: Preethi Wilson G, Gokulakrishnan R, Dhivya dhanasree S S, Sumanth BKM

Abstract: A virtual try-on system is an advanced AI-powered platform that allows users to visualize how clothing items would appear on their bodies without physically wearing them. These systems are transforming the way people shop online by offering a digital fitting room experience using computer vision, deep learning, and generative models. in recent years, the demand for online fashion experiences has increased, encouraging the development of systems like Style VTON, which not only allows users to try on clothes virtually but also supports multiple body poses and preserves personal identity and clothing details. By using a combination of input images (user photo, clothing image, and target pose), such systems generate a highly realistic image of the user wearing the desired outfit in a new posture

Published by: vikaspatanker

Authors: Sarvesh Bhandari, Rohan Wakchaure, Aarya Mahajan, Gauri Gadakh, Vaibhav Bhokare

Abstract: In the competitive and rapidly changing business world of today, organizations make greater use of business analytics to inform smart, sustainable choices. This paper discusses how analytics tools and techniques can help an organization enhance operational efficiency, time its forecasts better, and embrace strategies that could ensure long-term sustainability. Integrating descriptive analytics with diagnostic, predictive, and prescriptive analytics helps turn raw data into actionable insights for businesses to drive efficient resource utilization, better customer understanding, and strategic planning. The role of modern technologies, such as machine learning, business intelligence systems, and real-time dashboards, has also been discussed in enhancing the data-driven decisioning process. It also investigates how the application of business analytics can result in environmental, social, and economic sustainability by minimizing waste, optimizing operations, and encouraging responsible business operations. The study points out that based on the literature review and practical applications, there is a strong need for analytical competencies and a data-driven culture within organizations. The conclusions highlight that leveraging business analytics is an important pathway not only to attaining competitive advantage but also to sustainable and resilient business growth. This paper also emphasizes the importance of integrating sustainability goals into the analytical models that support balanced and responsible decision making. With the pressure by stakeholders, regulators, and consumers increasing in sustainability matters, being able to link performance metrics together with environmental and social indicators increasingly becomes a priority competency for business. Business analytics lets organizations assess the effects of their long-term decisions, measure sustainability performance, and helps organizations make decisions that not only benefit them but also align with global standards like ESG frameworks. By highlighting practical examples of emerging trends, the paper shows how analytics-driven insights empower organizations to innovate, reduce risks, and build sustainable value for all.

DOI: https://doi.org/10.5281/zenodo.18597175

Published by: vikaspatanker

Authors: Vishlesha Anil Habib, Vidya Gorakh Jagtap, Shrawani Ravindra Gaikwad, Nagraj Yashwant Kherud, Vaijayanti Pradip Kolhe

Abstract: The exponential growth of online platforms has enabled rapid dissemination of information, but it has also facilitated the widespread propagation of fake news. Fake news has negatively impacted political stability, public health, social harmony, and digital trust. This paper presents a comprehensive study and implementation of machine learning (ML) and Natural Language Processing (NLP)-based techniques for detecting fake news. The proposed system uses advanced text preprocessing, TF-IDF feature extraction, and multiple ML algorithms such as Logistic Regression, Support Vector Machine (SVM), Random Forest, and Naïve Bayes. Experimental results show that SVM achieves the highest accuracy of 94.8%, outperforming other models. This work demonstrates that combining linguistic features and machine learning provides a scalable and reliable approach to combat misinformation. Future enhancements include using transformer-based deep learning models and multilingual datasets

DOI: https://doi.org/10.5281/zenodo.18596939

Published by: vikaspatanker

Authors: Sachin Kumar

Abstract: The rapid evolution of wireless communication technologies has led to the emergence of sixth-generation (6G) networks, which aim to support ultra-low latency, massive connectivity, and intelligent network automation. One of the critical challenges in 6G is efficient mobility management due to highly dynamic user behavior, ultra-dense networks, and heterogeneous access technologies. Traditional mobility management schemes rely on reactive handover mechanisms that often result in increased latency, packet loss, and signaling overhead. To address these limitations, predictive mobility management has gained significant attention. This paper proposes the use of Long Short-Term Memory (LSTM) networks, a type of deep learning model well-suited for sequential data, to predict user mobility patterns in 6G networks. By leveraging historical mobility data, the LSTM-based approach enables proactive handover decisions, improved resource allocation, and enhanced Quality of Service (QoS). The paper discusses the architecture, working principle, advantages, and applicability of LSTM-based predictive mobility management in 6G environments, highlighting its potential to enable intelligent and autonomous network operations.

Published by: vikaspatanker

Authors: Yash Solunke, Ketan Bharambe, Nidhi Gandhi, Himani Suryawanshi, Khushi Raktate

Abstract: Sustainable irrigation is a critical component of modern agriculture due to increasing water scarcity, climate variability, and the need for precision resource management. Traditional irrigation systems, often based on fixed schedules or coarse environmental data, frequently lead to over-irrigation, under-irrigation, and inefficient water use. To address these limitations, this work introduces an AI-powered irrigation advisory framework that combines microclimate simulation, machine learning models, and real-time field-level sensing to generate accurate and adaptive water-use recommendations. The proposed system models localized microclimate parameters, including soil moisture, evapotranspiration, humidity flux, and temperature gradients, to provide more accurate short-term water demand estimates than traditional farm-level predictions. Machine learning algorithms continuously optimize the system, forecast crop-specific water needs, and dynamically identify patterns. To ensure robustness across diverse farming scenarios, the framework incorporates adaptive calibration mechanisms that adjust recommendations based on changing crop phenology and environmental conditions. We describe the implementation of this software-driven decision-support tool and its validation using both simulated and real-world agricultural datasets. Results demonstrate improved prediction reliability, a reduction in irrigation waste, and enhanced water-use efficiency compared to conventional scheduling methods. The proposed AI-powered sustainable irrigation advisor illustrates how microclimate-aware systems can advance next-generation smart agriculture, supporting productivity, environmental sustainability, and water conservation.

DOI: https://doi.org/10.5281/zenodo.18596198

Published by: vikaspatanker

Authors: Yash Solunke, Om Nikam, Shubham Chavan, Rutuja Raut, Pallavi Gulia

 

 

Abstract: BreathSafe is an innovative AIdriven system designed to monitor and diagnose respiratory conditions through breath analysis and real-time data processing. By leveraging machine learning algorithms on sensor data from wearable devices, BreathSafe enables early detection of diseases like COPD, asthma, and lung infections with over 90% accuracy in clinical trials. This paper presents the system's architecture, implementation, and evaluation for sustainable healthcare innovation.

DOI:

 

 

Published by: vikaspatanker

Authors: Piyush Kotkar, Pratik Halnor, Sakshi Kapse, Harshal Adhav, Atharva Dhawale

Abstract: Real-time foot traffic monitoring is now a key part of retail analytics, campus management, and smart surveillance. However, limitations in computing power make it hard to use heavy deep-learning models in low-power settings. This paper introduces a lightweight footfall counting system that uses YOLOv8n and YOLOv8s along with a centroid-based tracking method for effective ID persistence and directional counting. Experimental results indicate that YOLOv8n reaches 4.1 FPS on CPU-only systems with 98–99% ID stability, surpassing YOLOv8s in real-time performance. The system works well for embedded platforms, public monitoring, and budget-sensitive deployments.

DOI: https://doi.org/10.5281/zenodo.18596112

Published by: vikaspatanker

Authors: Mr. Viraj Kishor Chitte, Mr. Om Anant Aher, Mr. Darshan Santosh Bhandari, Mr. Sai Yogesh More, Mrs. Smita Manohar Dighe

Abstract: Neural networks, fuzzy logic, and expert systems are fundamental to the development of intelligent systems capable of addressing complex decision-making challenges across various domains. Neural networks, inspired by the structure of the human brain, demonstrate proficiency in pattern recognition, data classification, and high-accuracy prediction. Fuzzy logic facilitates reasoning under uncertainty, enabling systems to process imprecise inputs and generate responses that resemble human reasoning. Expert systems employ rule-based reasoning to emulate expert decision-making, delivering reliable solutions across healthcare, diagnostics, and industrial automation. This paper examines the underlying principles, strengths, limitations, and applications of these three artificial intelligence techniques. Through comparative analysis, it highlights their performance distinctions and unique contributions to intelligent problem-solving. Additionally, the study investigates the advantages of integrating these methods to create hybrid intelligent systems with improved adaptability, accuracy, and reliability. Such integrated approaches have the potential to advance AI-driven solutions in smart systems, real-time monitoring, and automated decision support.

DOI: https://doi.org/10.5281/zenodo.18596011

Published by: vikaspatanker

Authors: Kewal Manish Patel, Gaurav Tushar Kokate, Durvesh Amit Amale, Shubham Musmade, Atharva Gare

Abstract: Agriculture in India faces challenges such as unpredictable rainfall, improper irrigation planning, and inefficient use of water resources. To address these issues, this paper proposes a Smart Agriculture System that integrates Internet of Things (IoT) sensors with a lightweight Machine Learning model to optimize irrigation. The system collects real-time soil moisture, temperature, humidity, and light intensity data using low-cost sensors such as the soil moisture sensor and DHT11. The data is sent to a cloud platform through an ESP8266/NodeMCU microcontroller for monitoring. A simple ML model, such as Linear Regression or Decision Tree, predicts the required watering level based on sensor patterns. When moisture falls below the predicted threshold, the system automatically activates a water pump and sends an alert to the farmer’s mobile dashboard. The proposed solution reduces water wastage, increases crop health, and facilitates precision agriculture. This work demonstrates how IoT and ML together can support sustainable agricultural practices, contributing to UN Sustainable Development Goals (SDG-2 and SDG-12). The prototype is easy to implement, low-cost, and scalable for real-world applications.

DOI: https://doi.org/10.5281/zenodo.18595980

Published by: vikaspatanker

Authors: Jayashree Chava, Prasad Chavan, Dr. Pritish Vibhute

Abstract: Women’s safety continues to be a pressing concern globally, and timely access to help often determines the outcome of critical situations. With rapid advances in electronics and communication technology, there is growing potential to build practical tools that can offer support when it is needed most. This work presents a compact, AI-enabled wearable safety device developed specifically to assist women during emergencies. The device operates hands-free and relies on on-device voice recognition, implemented on an ESP32-S3 microcontroller trained using Edge Impulse. It uses Bluetooth Low Energy (BLE) to connect with a companion Android application. When the system recognizes the spoken keyword “Help! Help!” it functions entirely offline to activate the mobile app. The app then automatically fetches the user’s GPS location and sends an SOS alert to selected emergency contacts. It also uses the Google Places API to identify nearby police stations for quicker support. To strengthen post-incident reporting, the wearable includes an AI-based motion and image-capture module that records relevant visual evidence through its built-in camera. The prototype is designed to be power-efficient, affordable, and mindful of user privacy, making it suitable for both rural and urban environments. Overall, the proposed system shows how edge AI and IoT connectivity can be combined to create a practical and reliable personal-safety solution.

DOI: https://doi.org/10.5281/zenodo.18595926

Published by: vikaspatanker