Project Reports

Permanent URI for this community

http://172.16.1.132:4000/handle/123456789/14031

Browse

Now showing 1 - 20 of 92

Automated Office Meeting Summarization with NLP- Based Video Transciption and Speaker Diarization
(2024) Akash B - 1NH20AI006; C Sumukh - 1NH20AI019; Gowardhan Reddy V - 1NH20AI032; V Hashith 1NH20AI111
Automatic Speech Recognition (ASR) technology has revolutionized the way to capture and transcribe spoken words, enabling effective documentation and analysis of verbal communication. Despite significant advances, existing ASR systems for office meetings face persistent problems such as limited speaker diarization, insufficient punctuation recovery, and variable accuracy across languages. These limitations often result in transcripts that are difficult to follow and less useful for detailed meeting records. This project aims to develop an improved ASR system tailored specifically for office meeting environments. By utilizing advanced deep learning models and integrating features such as speaker diarization and punctuation recovery
Flashback.ai: A Study on Digital Second Brain
(2024) Mr. Syam Dev R S - 1NH20AI001; Abhishek Narsepalli Venkata sai - 1NH20AI002; Atman Mishra - 1NH20AI010
In an era defined by information abundance, the human capacit to capture, retian, and recall essential insights faces unprecedented challenges. The sheer volume of data encountered during online interactions and meetings often overwhelms our cognitive abilities, leading to the loss of valuable knowledge. Recognizing this critical gap, the Flashback.ai Project emerges as a revolutionary solution designed to bridge the divide between human memory limitation and expansive digital landscape.
Precision Farming with AI for Optimal Crop Yield and Health
(2024) A.VIGNESH REDDY - 1NH20AI007; D.GIRIHAS REDDY - 1NH20AI023; DAKSHA TM - 1NH20AI025; KUSHAL KULANDAIVELU - 1NH20AI051
In the Heart of India's Predominantly agrarian landscape, a groundbreaking web platform emerges as a beacon of transformative change. This Project, at the intersection of agriculture and technology, represents a monumental leap forward in empowering farmers and reshaping the country's economic landscape.
A Full-Fledged communicatin platform with Enhanced Accessibility and Functionality
(2024) Adhithya B N - 1NH20AI003; Chrish vinson Kunnankada (1NH20AI022); Pranit Prakash Prabhu - (1NH20AI078)
ProductGuard: Empowering Informed Choices for Safe Living
(2024) ADITYA S MANAKAR -1NH20AI005; ARSHAD PASHA - 1NH20AI009; JIYA ANN ELIAS - 1NH20AI037; KARTIKEY TIWARI - 1NH20AI044
In a landscape where consumer decisions significantly influence personal health and wellbeing, the necessity for informed choices regarding product safety is paramount. This project introduces ProductGuard, a mobile application designed to empower users in making conscientious and informed decisions about skincare and personal care products. The project integrates four core components aimed at assisting users in evaluating product safety, comparing ingredients, providing personalized recommendations based on skin type, and offering expert guidance through a chatbot interface. The first component employs advanced image analysis technology utilizing a dataset extracted from the Environmental Working Group (EWG) website, enabling users to capture product ingredient images and receive detailed safety analyses.
Anemia Detection Using Machine Learning
(2024) A Chetu Chandhan - 1NH20AI004; K Lingeswar Balaji - 1NH20AI040; K Datta Ram Vivek - 1NH20AI046; P Sanjay reddy - 1NH20AI143
This study investigates the feasibility of using eye datasets for predictive modeling. Three machine learning algorithms, decision tree, random forest, and XGBoost, were employed to classify individuals with and without anemia. A comprehensive dataset of eye images was obtained and preprocessed to capture key characteristics like color variations, textural patterns, and structural details. These features were then fed into the respective algorithms to construct predictive models. Evaluation metrics, including accuracy, revealed promising performance from all three models in identifying anemia based on eye imagery. Notably, the XGBoost algorithm achieved the highest accuracy, followed by random forest and decision tree. These findings suggest that eye images hold significant potential as a non-invasive and cost-effective tool for early anemia detection. The developed machine learning models utilizing decision tree, random forest, and XGBoost offer a promising avenue for further research and development in this area.
Web-Based Real-Time Child Surveillance System
(2024) NHCE
In an increasingly busy world, guardians often face challenges in ensuring the safety and well-being of their children while away from home. To address this issue, we have developed a "Web-Based Real-Time Child Surveillance System" utilizing Python and Convolutional Neural Networks (CNNs) for real-time monitoring. With the use of a web application, guardians will be able to remotely watch their children's actions and receive real-time notifications. The monitoring system recognizes and categorizes a range of child behaviors, including walking, running, sitting, and falling. With the use of real-time processing and sophisticated machine-learning techniques, our technology provides parents and guardians with a dependable and effective way to monitor their kids while they are away.
Deepfake Detection Using Multi Modal Approach
(2024) NHCE
Deepfake technology is a serious danger to the accuracy of information shared online in the age of digital communication. This artificial intelligence (AI) produced videos, which may accurately portray people saying or doing things they never did, have serious ramifications for digital media credibility, human rights, and public debate. Advanced techniques for deepfake detection are required due to their potential misuse for espionage, manipulation, coercion, and harassment. In order to overcome this difficulty, we have created a deepfake video detector by utilizing CNNs' capabilities. Our method examines video frames for minute discrepancies that are characteristic of deepfake footage, making use of CNN's powerful feature extraction capabilities. Our methodology provides a potential remedy for by concentrating on temporal irregularities and pixel-level differences that are frequently undetectable to the human sight. This effort not only advances technology in the battle against digital disinformation, but it also emphasizes how crucial cross-sector cooperation is to preserving the integrity of online media. Our results shed light on the direction of future studies and advancements in the industry and demonstrate how important sophisticated machine learning methods are to preserving the security and legitimacy of digital interactions. It's important to acknowledge, however, that the fight against deepfakes is an ongoing arms race. As deepfake creators develop more sophisticated techniques, so too must deepfake detectors. This necessitates continuous improvement of detection algorithms, collaboration between researchers and tech companies, and public awareness campaigns to equip users with critical thinking skills to spot potential deepfakes.
Multiple Disease Detection
(2024) NHCE
The rapid advancement of deep learning techniques has significantly impacted the field of medical diagnostics, offering promising solutions for multiple disease detection. This study explores the application of deep learning algorithms to accurately identify and diagnose multiple diseases from medical imaging and other healthcare data. The proposed system leverages convolutional neural networks (CNNs) and recurrent neural networks (RNNs), combined with advanced data preprocessing and augmentation techniques, to enhance diagnostic accuracy and efficiency.
Stress Analysis and Prediction using Machine Learning with EEG
(2024) NHCE
Stress, which is known to be the underlying cause of many of mental health disorders, arises from various of sources that have a clearly harmful effects on health. The consequences of stress are most noticeable in the life of a working professional who must handle the demands of increased management expectations, time management limitations, and family obligations. Proactive stress management is crucial since ignoring stress for a long time increases the likelihood of developing anxiety and depression. Physiological characteristic is very essential for diagnosing stress-related conditions and provide important information about the complex relationship between mental and physical health. The goal of this study is to investigate stress using EEG data, which is recognized for its reliability, accuracy, and precision. Advanced (ML) machine learning models, such as SVM, RF, Decision Tree (DT), and KNN, are implemented based on the intrinsic compatibility between stress signals and EEG data.
Automation of Statistical Aid to Data Engineering
(2024) New Horizon College of Engineering
In the realm of data cleaning, statistics plays a pivotal role in elucidating patterns, identifying anomalies, and guiding the selection of optimal strategies for enhancing data quality, statistics provides a systematic framework for analyzing and interpreting data, offering valuable insights that are instrumental insight that are instrumental in the data cleaning process.
Personalized dietary guidance with AI
(2024) New Horizon College of Engineering
In today's world, that has an abundance of food options, customers frequently encounter difficulties in making knowledgeable nutritional choices. Sorting through the numerous choices that are available and each claiming to be the greatest for your health can be rather overwhelming. Acknowledging this difficulty, our project seeks to transform the relationship between nutrition and technology by focusing the requirements and welfare of the user. There is a shift in mindset in which people are empowered by technology to make personalised, well-informed decisions about their nutrition and diet in addition to receiving information.
AI- Trip Advisor [Smart Travel Itinerary planner]
(2024) DAIVIK SUCHIT -1NH20AI024, JEEVAN C B - 1NH20AI036 , M DATHRI VENKAT REDDY -1NH20AI054, LIKITH M -1NH20AI134
State-of-the-art technologies—Language Models (LLMs) in particular—are used to revolutionize conventional travel planning procedures. The "AI Trip Advisor" web application uses cutting-edge LLMs to provide consumers with a smooth and customized travel schedule. The system takes into account a number of variables, such as the user's location, preferred destinations, travel dates, financial limits, traveling companions, and customized instructions, like adding rest days or certain activities. Intelligent Flight Recommendations, wherein the system proposes ideal flight options based on user preferences, including cost, travel time, and layovers, is one of the application's key features. Budgetary restrictions, user preferences, and accessibility to well-known sites are all taken into account when providing accommodation suggestions. Users can modify their trip schedules to suit their own interests and needs using customizable itineraries. Weather and Best Times to Visit recommendations optimize the overall travel experience by considering weather conditions. The application ensures a fair allocation of costs among trip companions by offering Expense Splitting for group travelers, which tackles the financial part of travel as well. The domain is distinguished by the notable progress in incorporating artificial intelligence into the process of travel planning. The program offers a clever and user-friendly solution while streamlining the planning process to save users time and effort. This novel strategy represents a revolutionary change in the way travel plans are tailored and optimized, and it is in line with the needs of contemporary travelers.is started, the writer must exercise ingenuity in figuring out how to fit the paragraph into their work.
AI-Powered Video Conferencing App with Real-time Speech-to-Speech Machine Translation
(2024) NHCE
In the contemporary landscape of communication, video conferencing has emerged as an indispensable tool for global interaction, revolutionizing the way individuals and organizations connect across vast distances. The proliferation of video conferencing platforms has enabled seamless collaboration, transcending geographical boundaries and fostering real-time dialogue among users. This technological advancement has facilitated everything from business meetings and educational sessions to social gatherings and telehealth consultations, making face-to-face interaction possible regardless of physical location. However, despite these advancements, a significant barrier persists: language differences. These linguistic obstacles can impede the free flow of ideas, hinder effective discourse, and ultimately limit the inclusivity and efficiency of virtual communication. Language barriers can lead to misunderstandings, reduce engagement, and prevent participants from fully contributing to discussions. Addressing this challenge is crucial for enhancing the accessibility and utility of video conferencing technologies, ensuring that they can truly serve a diverse, global audience by bridging the communication gap and fostering a more inclusive virtual environment.
Graph based Retrieval Augmentation for Education and Learning
(2024) DHARSHAN AN : 1NH20AI026; G SAI CHARAN: 1NH20AI031; JAYAVIBHAV NK: 1NH20AI035; K MANISH : 1NH20AI038
Graph-based knowledge representation has emerged as a powerful technique for organizing and structuring information in a way that captures rich semantic relationships. By representing knowledge as a network of interconnected nodes and edges, graph structures allow for efficient encoding of complex conceptual associations and hierarchies. This approach has significant implications for information retrieval and generation tasks, particularly in the realm of retrieval augmented generation.
Platform epoch
(NHCE, 2024) AMOGH.S.KUMAR-1NH21CS280 AARAMBH JAISWAL -1NH21CS283 T SAI SREE CHARAN -1NH21AI109 VIBHA JOSHI -1NH21AI117
Children with Autism Spectrum Disorder (ASD) and Attention Deficit Hyperactivity Disorder (ADHD) often face unique developmental challenges that demand tailored interventions and educational support. This project, "Platform Epoch” seeks to bridge the gap between advanced diagnostic tools and personalized learning systems. By leveraging the power of digital identity and learning analytics, the platform provides an integrated solution to support early detection, informed decision-making, and customized educational pathways for children with special needs. At the core of the platform lies a machine learning model that computes an autism risk score (A_Score) by analyzing behavioral and demographic data collected through structured input forms. The model employs advanced neural network architectures to classify risk levels with precision, ensuring timely and accurate diagnostics. Complementing this functionality is a dynamic learning ecosystem that uses learning analytics to track progress, identify strengths and weaknesses, and provide actionable recommendations for parents and educators. By offering a robust backend powered by Flask APIs and secure data handling mechanisms, the platform ensures real-time processing and reliability. The platform is designed with accessibility and inclusivity in mind, emphasizing features like intuitive dashboards, responsive interfaces, and multilingual support. A unique digital identity system links parent and child accounts, facilitating personalized interactions while ensuring data privacy. Additionally, the system incorporates adaptive learning pathways tailored to the child’s cognitive and behavioral profiles, enabling gradual improvement and sustained engagement. Visualization tools such as Matplotlib and Plotly enhance user experience by presenting insights in an interactive and comprehensible manner.
Design and Fabrication of Smart Wheelchair-Cum-Bed Equipped with Health Monitoring System
(NHCE, 2025) SURESH A 1NH22ME412 BHAGATH MANJUNATH S 1NH22ME402 SWASTITIRTHA DASH 1NH21AI108 SAI KRISHNA KM 1NH22AI409
This project focuses on the design and development of a smart Wheelchair-Cum-Bed. capable of seamlesaly transforming into either a wheelchair or a bed at the press of a button. The system incorporates a rack-and-pinion mechanism for smooth transition between modes, ensuring ease of use for individuals with mobility challenges. Additionally, the wheelchair can be moved from one location to another with a simple button press. Integrated health monitoring features, including sensors for heart rate, pulse, humidity, and ECG, provide real-time health data to enhance the user's well-being. This innovative design aims to offer both mobility assistance and healthcare support, improving comfort, independence, and affordability for users
Plant disease detection and diagnosis
(NHCE, 2025) CHANDANA KS 1NH22CS404 MAHARATHI GN 1NH22CS410 NITHISH SOMANI 1NH21AI100 ABHILASH G M 1NH22AI400
The "Plant Disease Detection and Diagnosis using Deep Learning" project is designed to provide an innovative and effective solution to ansist farmers in identifying and diagnosing plant diseases in real-time. The system leverages TensorFlow Lite, a lightweight framework optimized for mobile and edge devices, to deploy a deep learning model trained on a dataset of 1.5k images. These images represent seven plant disease classes, including rust, mold, blight, rot, powdery mildew, scab, and spot. By using computer vision techniques combined with Mediapipe for real-time detection, the system can analyze plant images captured from mobile devices or cameras, offering instant feedback to farmers regarding the presence of disease. The deep learning model has been specifically designed to run efficiently on resource-constrained devices, ensuring that farmers can easily deploy the solution in the field. The backend of the system is powered by Flask, while the frontend imerface is built using HTML, CSS, and JavaScript to provide a simple, user-friendly platform. The web interface allows farmers to upload images of their plants, where the system analyzes the data, diagnoses the disease, and recommends appropriate remedies. This real-time diagno sis helps farmers take immediate corrective actions to prevent the spread of diseases, reducing potential crop losses and improving productivity. In addition to diagnosis, one of the key features of this system is its ability to support multiple languages. Using the Google Translate API, the system can translate disease diagnosis and remedial advice into several regional languages, including Kannada, Hindi, Malayalam, Telugu, and Tamil. This multilingual feature ensures that the system is accessible to farmers in different linguistic regions, empowering them to understand the diagnosis and suggested remedies in a language they are familiar with, thereby bridging communication gaps in agriculture.
Adaptable Arms: innovations in flexible robotic manipulation
(NHCE, 2025) SURESH R 1NH22A1410 VARUN MS 1NH22A1411 SANGAMESH POLICE PATIL 1NH22ME411 SUDEEP 1NH21ME073
The problem definition of flexible robotic arm movement in the agriculture field centres around the need for automation solutions that can handle the diverse, unpredictable, and often delicate tasks involved in modern farming. Agriculture presents unique challenges that traditional rigid robotic arms struggle to address due to the variability of the environment, crop types, terrain conditions, and the need for precision in handling perishable goods. Flexible robotic arms offer a promising solution to these challenges, but their design and deployment come with several complex problems that need to be solved.
Real-time fire and smoke detection and retardation
(NHCE, 2025) HEMANTH KUMAR REDDY Ρ 1ΝΗ21Α1127 RAHUL D 1NH21A1128 MANU M 1NH21EC090 NANDEESH GOWDA B 1NH21EC102
Fire and smoke detection systems play a vital role in ensuring safety and reducing losses caused by fire-related incidents. However, conventional systems that rely on heat and smoke sensors are often susceptible to false alarms and may fail to detect fires at their earliest stages. This paper presents a modern, cost-effective, and intelligent approach to fire and smoke detection, utilizing the ESP32-CAM microcontroller combined with a Convolutional Neural Network (CNN). The ESP32-CAM, a compact and affordable lot-enabled microcontroller with an integrated camera, serves as a reliable platform for real-time image capture and processing. When paired with CNNs, a cutting-edge deep learning architecture for image recognition, the system provides a robust and highly accurate solution for detecting fire and smoke. By integrating lot and Al technologies, the proposed system bridges the gap between affordability and advanced functionality, ensuring timely and precise alerts The system architecture comprises three core components: the ESP32-CAM for image acquisition, the CNN for image-based analysis, and a communication interface for transmitting alerts. The ESP32-CAM continuously captures images or video streams, which are then preprocessed to enhance detection accuracy. Preprocessing techniques, such as resizing, normalization, and noise reduction, prepare the visual inputs for the CNN. The trained CNN identifies critical features in the data to detect fire and smoke reliably. At the heart of the system lies the CNN, optimized specifically for deployment on resource-constrained devices. Its lightweight design ensures compatibility with the ESP32-CAM's limited computational capabilities while maintaining high accuracy. Optimization techniques like model pruning, quantization, and transfer learning help reduce computational demands without compromising performance.

Browse

Browsing Project Reports by Issue Date

Results Per Page

Sort Options