2024-25

Permanent URI for this collection

http://192.168.75.5:4000/handle/123456789/18850

Browse

Now showing 1 - 20 of 78

Hardware implementation of smart wearable device with multi sensor integration for search and rescue operations
(NHCE, 2025) Aniruddh G Bharadwaj 1NH21EC017 Gagan M 1NH21EC056 Chandrahas Sai R 1NH21EE025 Manjunath V 1NH21EE055
Search and rescue operations are inherently risky, often taking place in unpredictable and hazardous environments. Personnel face numerous dangers, including physical injuries, exposure to toxic substances, and disorientation. Traditional methods for tracking personnel and monitoring environmental conditions during these operations rely on bulky equipment, intermittent communication, and delayed data transmission, hindering effective decision-making and jeopardizing rescuer safety. This thesis presents the development and evaluation of a novel "Smart Wearable Device" designed to address these limitations and revolutionize search and rescue operations by enhancing safety, efficiency, and situational awareness. This device integrates a comprehensive suite of sensors into a compact, wearable unit, providing real-time data on both the rescuer's physiological condition and the surrounding environment. Physiological monitoring includes heart rate and galvanic skin response (GSR) sensors, offering insights into the wearer's physical exertion and stress levels. Environmental monitoring capabilities encompass gas detection (MQ135, MQ2) for identifying hazardous substances, particulate matter measurement (PMS7003) to assess air quality, and temperature sensing (Dallas DS18B20) for detecting potential hypothermia or hyperthermia risks. These physiological and environmental data points contribute to a comprehensive understanding of the rescuer's well-being and the challenges posed by the operational environment. By precise location tracking using GPS and MEMS sensors, enabling continuous monitoring of the rescuer's position and movement. This real-time location data is relayed wirelessly via a GSM module, facilitating immediate response in case of emergencies or deviations from the planned route. A dedicated SOS button allows rescuers to instantly transmit distress signals to a central command unit, ensuring prompt assistance in critical situations. All sensor data is processed by an Arduino Mega and transmitted to a cloud platform via a NodeMCU module, enabling remote monitoring and analysis by the command center. This cloud connectivity allows for real-time visualization of the rescuer's location, physiological status, and surrounding environmental conditions, providing a comprehensive overview of the situation and facilitating informed decision-making. By combining real-time data acquisition, wireless communication, and an intuitive interface, this Smart Wearable Device aims to bridge the information gap that often hinders effective search and rescue operations.This Smart Wearable Device has the potential to revolutionize search and rescue operations by empowering rescuers with real-time information, enhancing communication capabilities, and ultimately, increasing the safety and effectiveness of those who dedicate themselves to saving lives.
Multilingual Real-Time Voice Translator with Emotion Detection and Web Interface
(NHCE, 2025) Mithun S 1NH21EC093; Monisha M 1NH21EC097; Mudunuri Aditya Varma 1NH21EC099; Kamal Deep U 1NH21ME032
Access In an increasingly interconnected world, seamless communication across linguistic boundaries has become a must. This paper details the development and progression of a Multilingual Real-Time Voice Translator designed to bridge language barriers and foster global inclusiveness. The system was initially developed using Python libraries, including google translate, and speech recognizer, to provide basic real-time speech and text translation capabilities. It has transformed into one having a web interface and emotion detection for better usability and understanding. This makes the translator a powerful tool for multilingual communication and emotional insight. Accessibility and user experience were highly improved by the web-based interface. Modern technologies built this system to be easily used on various devices with seamless interactions. Real-time speech-to-text, multilingual selection, and audio playback make the system versatile and user-friendly. An emotion detection feature using a Transformers-based deep learning model has been added to enhance communication. It analyzes vocal intonations and textual cues to identify emotions like happiness, sadness, anger, and neutrality. This is particularly useful in mental health, customer service, and cross-cultural communication, where emotional context matters. The translator now covers more languages and dialects. Context-aware algorithms boost translation accuracy, allowing the system to manage idioms and complex structures. These improvements ensure reliable communication across various languages. The Multilingual Real-Time Voice Translator is useful in business, education, healthcare, and social interactions. It aids communication in international meetings, supports language learning, and enhances doctor-patient conversations. Its emotion detection feature allows professionals to assess emotional states, promoting understanding and inclusivity.
Creating an adaptive robot for personalized Health monitoring and assistances
(NHCE, 2025) Gnanesh H Nayak 1NH21EC059 Siddeshwar S Kode 1NH21EC149 Sai Pranathi T 1NH21CS207 Sai Rishitha Ambati 1NH21CS209
Based on the latest robotics trends updated in 2019 by the ROBO Global Robotics & Automation Index, healthcare now constitutes 10% of the index, focusing on robotics for surgical guidance, laboratory automation, genomics, and Al applications in healthcare. The sector showed a robust return of over 21% in the first 11 months of 2018 [1]-[3]. This highlights the increasing role of robots in healthcare, particularly in enhancing patient care and efficiency. Advanced countries like Japan and Europe are increasingly adopting robots, which often outperform humans in certain tasks
SMARTNAV
(NHCE, 2025) Rishab Swaroop 1NH21EC129; Ramakrishna U 1NH21EC128; Aditya Sachin Sawant 1NH21IS008; Ishan Sanjeev Paul William 1NH21IS065
For individuals with visual impairments, independent navigation can be a daunting and often unsafe task, heavily impacting their mobility and confidence. SmartNav is a solution designed to address these challenges, providing an intelligent assistive device that combines cutting-edge technologies to offer practical, reliable, and user-friendly navigation support. By integrating real-time object detection, spoken feedback, and location guidance, SmartNav enables users to move more safely and independently in their environment. The system utilizes a Raspberry Pi and a USB camera to detect obstacles in real-time. The camera processes visual input, identifying potential hazards, while the Raspberry Pi provides the computational power needed to process the data and generate timely feedback. A GPS module ensures that location-based guidance is available, helping users navigate longer routes or unfamiliar areas with confidence. To make the device even more user-focused, spoken instructions are used to communicate vital navigation information. These instructions ensure clarity and reduce the cognitive load on users, allowing them to focus on their surroundings rather than interpreting complex signals. One of the standout features of SmartNav is its chatbot, which enhances the user experience by offering an interactive layer of support. This conversational companion not only provides additional assistance during navigation but also helps reduce feelings of isolation, making the device a true companion for visually impaired individuals. SmartNav is designed with scalability and affordability in mind. Its modular structure makes it easy to incorporate future upgrades, such as advanced object detection algorithms, gesture-based controls, and IoT connectivity for remote monitoring and data sharing. The project’s focus on accessibility ensures that it remains a cost-effective option for individuals and organizations looking to improve the quality of life for visually impaired users
Neuro-Voice
(NHCE, 2025) Akshay Gowda S 1NH21EC012 Raghava Reddy N 1NH21EC125 Koyal Chadresh 1NH21AI045 Tharnish K M 1NH21AI137
Many of the existing machine learning models for health care analysis are concentrating on one disease per analysis. Like one analysis if for diabetes analysis, one for cancer analysis, one for skin diseases like that. There is no common system where one analysis can perform more than one disease prediction. In this article proposing a system which used to predict multiple diseases by using Flask API. In this article used to analyse Diabetes analysis, Diabetes Retinopathy analysis, Heart disease and breast cancer analysis. Later other diseases like skin diseases, fever analysis and many more diseases can be included. To implement multiple disease analysis used machine learning algorithms, tensorflow and Flask API. Python pickling is used to save the model behaviour and python unpickling is used to load the pickle file whenever required. The importance of this article analysis in while analysing the diseases all the parameters which causes the disease is included so it possible to detect the maximum effects which the disease will cause. For example for diabetes analysis in many existing systems considered few parameters like age, sex, bmi, insulin, glucose, blood pressure, diabetes pedigree function, pregnancies, considered in addition to age, sex, bmi, insulin, glucose, blood pressure, diabetes pedigree function, pregnancies included serum creatinine, potassium, GlasgowComaScale, heart rate/pulse Rate, respiration rate, body temperature, low density lipoprotein (LDL), high density lipoprotein (HDL), TG (Triglycerides). Final models behaviour will be saved as python pickle file, Flask API is designed. When user accessing this API, the user has to send the parameters of the disease along with disease name. Flask API will invoke the corresponding model and returns the status of the patient. The importance of this analysis to analyse the maximum diseases, so that to monitor the patient's condition and warn the patients in advance to decrease mortality ratio.
loT Induced Vertical Farming
(NHCE, 2025) Gagan P 1NH21EC057; Akhil aik G 1NH21EC011
The integration of Internet of Things (IoT) technologies into agriculture has emerged as a transformative solution to address pressing global issues such as food security, resource management, and urban growth. Among these innovations, IoT-driven vertical farming is revolutionizing urban food production by enabling the cultivation of crops in vertically stacked layers, often within controlled indoor spaces like warehouses or buildings. This approach harnesses a network of interconnected sensors, data analytics, and automation systems to optimize growing conditions for plants, making it an ideal solution for space-limited urban areas. Vertical farming itself is a highly efficient method that maximizes crop yield per square meter, which is particularly valuable in cities where traditional farmland is scarce. By incorporating IoT technologies such as environmental sensors, humidity controllers, nutrient delivery systems, and smart lighting, vertical farms can be continuously monitored and adjusted to ensure optimal conditions for plant growth. Real-time data on parameters such as temperature, humidity, light levels, CO2 concentration, and nutrient balance enable precise control over irrigation, lighting, and climate, resulting in improved efficiency, minimized waste, and consistent crop quality. Beyond simple automation, IoT in vertical farming offers predictive analytics that empower farmers to make data-driven decisions on resource allocation and crop management. This not only increases productivity but also minimizes the environmental impact of agriculture by reducing water and energy consumption. Additionally, IoT systems enable remote monitoring, allowing farmers to manage operations from any location, further enhancing the flexibility and scalability of the farming model. The incorporation of IoT into vertical farming marks a significant step toward sustainable urban agriculture, especially in regions where land, water, and other resources are limited. Despite its potential, the widespread adoption of IoT-enabled vertical farming faces challenges such as high upfront costs, technical integration issues, and concerns regarding data security. Nonetheless, this technology holds great promise for decentralizing food production, making it more resilient to climate change and disruptions in global supply chains. In conclusion, IoT-enabled vertical farming is a promising innovation at the intersection of technology, sustainability, and urban development. As the technology advances, it has the potential to reshape global food production, contribute to the growth of smart cities, and support more sustainable agricultural practices worldwide.
Anti-sleep alarm for Accident prevention
(NHCE, 2025) Dibyanshu P 1NH21EC046; Mahi Prajwal 1NH21EC087; S Deep 1NH21IS127; Naveen A 1NH22IS407
Drowsy driving is one of the most leading causes of global road accidents that cause heavy casualties and loss. The Anti-Sleep Alarm with Engine-Lock system is specifically designed as a smart safety tool to reduce sleep-related accidents in the future. The system applies advanced technologies such as an eyeblink sensor, an Arduino Nano microcontroller, and multi-layered responses to monitor drowsiness levels in drivers real-time. The main functionality is to identify the physiological signs of fatigue, including prolonged eye closure, head nodding, and yawning patterns. Once these signs are detected, the system activates the escalating response. Initially, it issues warnings with audible alarms and visual signals to awake the driver. The driver's unconscious state will engage the engine-lock mechanism that would gradually bring down the vehicle's speed, slowing it to a controlled stop while protecting the safety of the driver, passengers, and other road users. This project focuses on easy integration into any type of vehicle, using a scalable design to minimize modifications and support diverse applications, from private cars to commercial fleets. Its non-intrusive nature and user-friendly interface enhance its adoption potential, and its energy-efficient design ensures long-lasting operation. Beyond immediate safety factors, it matches modern trends in automobiles, which call for intelligent and responsive technologies. This system will solve the basic causes of drowsy driving with incorporated high-end automation. It shall thereby form a bench mark for future road safety innovations, reduce costs associated with healthcare and repair following accidents, and boost people's confidence in preventive measures. The project is thus one step forward to safer roads, with advanced technology to save lives and rewrite the rules of driving safety.
Next generation industrial safety - smart helmet
(NHCE, 2025) Brunda v 1NH21EC032 Chiranthana M Reddy 1NH21EC038 Jefferson Randy Dcosta 1NH21ME026 Kenneth Aaron Fernandez 1NH21ME034
The Smart Helmet for Industrial Applications is a groundbreaking innovation aimed at redefining workplace safety and efficiency in hazardous environments. In industries such as construction, mining, and manufacturing, workers are often exposed to a myriad of risks, including toxic gases, extreme temperatures, and physical fatigue. This project introduces a cutting-edge safety solution that integrates Internet of Things (IoT) technologies, advanced sensors, and real-time data analytics to address these challenges and ensure the well-being of industrial workers. The primary objective of the smart helmet is to provide comprehensive safety monitoring and instant feedback to both the user and central safety management systems. By incorporating sensors that measure environmental conditions-such as temperature, humidity, and gas concentrations-the helmet can detect potentially dangerous situations in real time. It also tracks worker vitals, including heart rate and fatigue levels, offering an added layer of protection by identifying early signs of physical strain or health risks. These capabilities make the smart helmet an indispensable tool for minimizing workplace hazards and preventing accidents. A key feature of this innovative device is its ability to deliver immediate alerts through audible, visual, and haptic signals when hazardous conditions are detected. For example, the presence of toxic gases or a sudden rise in temperature triggers the helmet's alert mechanisms, allowing workers to take swift action to safeguard themselves. Additionally, the integration of GPS technology enables real-time location tracking, which is crucial for managing emergency responses and ensuring efficient workforce coordination. Supervisors can monitor the location and safety status of workers, thereby optimizing operational workflows and enhancing overall productivity.
Real time translation of Indian languages
(NHCE, 2025) G harshath Kumar 1NH21CS284 T Manoj 1NH21CS293 B. Shashank 1NH21EC031 D Harshavardhan 1NH21EC041
The translation of Indian languages is very important in many different ways such as business expansion internationally and nationally, health care departments, and communicational purposes. This integrated tool offers a subtle experience by providing the multi lingual conversion of Indian languages. It converts different dialects of Indian languages by breaking the barrier between users who struggle to communicate. It fosters efficient multi -lingual communication and understanding. It is an invaluable asset for the users who are looking for rapid and accurate language translations. It signifies the groundbreaking barriers of the multi-language barrier between the users who knows many languages by accommodating to provide many languages. This offers us many languages which turns out into Telugu, Tamil, Marathi, Kannada, Spanish, French, Japanese, and Russian which are very important languages around Indian territory and our neighboring countries. This integrated tool also offers audio message for the users who are visually impaired. It extracts the text from the images and it translates the information to which language is demanded by user.
Intellectus: personalized virtual teaching assistant
(NHCE, 2025) Nithik C Reddy 1NH21EC112; Niven Nathaniel D 1NH21EC113; G Bala Sai Reddy 1NH21A1033
The Intellectus project is an effort in the pioneering attempt to change the education landscape by the infusion of artificial intelligence, machine learning, and natural language processing. It was developed as a smart virtual teaching assistant, focusing on critical issues of conventional education such as automating the retrieval of information, personalization of learning, and creating a culture of inclusiveness. By using cutting-edge technology such as Bidirectional Encoder Representations from Transformers and Automatic Speech Recognition (ASR), this system delivers real-time contextual answers to academic questions presented in both text and audio forms. This project is designed to create a user-friendly, web-based interface that accommodates diverse learning preferences and enhances accessibility for students and educators, including those with visual or auditory impairments. The advanced NLP techniques and content extraction tools such as PyMuPDF help Intellectus process vast academic resources efficiently and transform them into actionable knowledge. Moreover, the system is designed with a continuous feedback mechanism to ensure its evolution in response to dynamic user needs and emerging educational trends. By streamlining routine tasks for educators and promoting interactive engagement for students, Intellectus establishes itself as a versatile and scalable solution across varied educational settings. Its modular architecture supports adaptability across disciplines, catering to both primary education and advanced professional training programs. In doing so, Intellectus transcends the limitations of conventional pedagogical methods, creating an inclusive, dynamic, and effective learning environment that empowers users to achieve their full potential.
Waste segregation detector
(NHCE, 2025) Bhoomika Srinivas 1NH21EC029 C Yashvitha 1NH21EC034 Harish Rajesh 1NH21CS099 Jaishankar S 1NH21CS105
In this recent world, urbanization has increased tremendously. At the same phase, there is increasing amount of in waste production. Waste management has been a crucial issue to be considered. This report is a different way to achieve this good cause. In this report, smart bin is built on a microcontroller-based platform Arduino - Uno board, which is interfaced with Ultrasonic sensor. It will stop overflowing of dustbins along roadsides and localities as smart Dustbins are managed in real time. Once these smart bins are implemented on a large scale by replacing the traditional bins, the waste can be quickly managed to its efficient level as it avoids unnecessary lumping of wastes on roadside. Foul smell from these rotten wastes that remain untreated for a long time, due to negligence of authorities and carelessness of public may lead to long term problems. Breeding of insects and mosquitoes can create nuisance around promoting unclean environment. This may even cause dreadful diseases. The goal of this project is to keep our environment clean. It also aims at creating a clean as well as green environment.
Dynamic wireless charging for electronic vehicles for smart transportation system
(NHCE, 2025) Deepak M 1NH21C5066; Ketankumar M 1NH2105131; Bhagyashree Chikkeri 1NH21EC028; Vibha Adur 1NH21EC178
Electric vehicles (EVs) are more popular due to advanced battery technology and concerns about the environment. However, one main challenge holding back their widespread use is "range anxiety." This is the fear that an EV might run out of battery before reaching a charging station. To solve this problem and make electric mobility more convenient, a new idea focuses on creating "intelligent roads" equipped with technology for wireless charging while vehicles are on the move, called dynamic wireless charging. The study explores two key methods for this wireless charging technology: inductive charging and magnetic concrete. Inductive charging works by using embedded coils in the road surface to transfer energy wirelessly to a receiver coil installed under the EV. Magnetic concrete, involves mixing special magnetic materials in the concrete used for roads. This type of road can interact with the EV’s receiver coil, turning the surface into a charging system. Although the concept is promising, there are many technical challenges that should be addressed. The main concern is ensuring efficiency—minimizing energy loss during the transfer process so that the EV receives enough power while driving. Another critical issue is safety, particularly with regard to electromagnetic fields, which could pose risks to nearby people, animals, or other vehicles. Additionally, ensuring that the infrastructure can avoid interference with metal objects on the road is a priority. The study also considers the economic feasibility of building and maintaining such intelligent road systems, as the implementation cost could be significant. Dynamic wireless charging offers many advantages for EVs. First and foremost, it can reduce range limit issues, which is huge barrier to EV adoption. Drivers will no longer need to worry about stopping at charging stations frequently; instead, their cars can recharge while they drive, much like refueling traditional gasoline-powered vehicles. This convenience can make the experience of owning and driving an EV much more appealing. Moreover, with more people likely to switch to EVs, there would be a positive environmental impact, as society’s dependence on fossil fuels would decrease, promotes cleaner air and reduces greenhouse gas emissions. To conclude, dynamic wireless charging through intelligent roads is a promising solution to range anxiety and can promote the transition to green transportation systems. However, further research and development are crucial to overcoming the technical and economic challenges involved. If successful, this revolutionary technology could transform the thought about transportation, making it more efficient, sustainable, and convenient for everyone. In the future, these charging highways could play a main role in the evolution of electric vehicle technology, helping to create a smarter and more environmentally friendly mobility ecosystem.
IoT based Heart Rate Monitoring and Heart Attack
(NHCE, 2025) Chetan L D 1NH21EE027 Mohaka R 1NH21EC094 Gunturu Chiranjeevi 1NH21EC062
The Internet of Things (IoT) has revolutionized healthcare by enabling real-time monitoring and remote accessibility. This project presents an IoT-based system for heart rate monitoring and heart attack detection, addressing the critical need for timely cardiovascular care. Cardiovascular diseases are a leading global cause of death, and early intervention can save lives. The system employs a pulse oximeter sensor to measure heart rate and oxygen saturation (SpO2), with data processed by an Arduino Uno microcontroller and displayed on an LCD screen for local monitoring. The integration of a NodeMCU ESP8266 WiFi module facilitates remote data transmission to a cloud platform, enabling caregivers and healthcare professionals to monitor patients from anywhere. Immediate alerts via LED indicators and a buzzer ensure prompt action during emergencies. Additionally, an L298N motor driver allows control of external actuators, enhancing its adaptability for therapeutic interventions. This system is particularly beneficial for remote health diagnostics and elderly care, reducing the need for frequent hospital visits while empowering users to proactively manage their health. In clinical settings, it aids continuous patient monitoring, lessening the workload on medical staff. Its cost-effective and compact design makes it suitable for wearable applications. Future enhancements include adding sensors for blood pressure, temperature, and ECG readings, implementing machine learning for predictive analysis, and optimizing energy efficiency for portability. Ensuring robust data security and developing user-friendly interfaces will further enhance its utility. By leveraging IoT, this system offers a scalable, accessible, and proactive approach to cardiovascular health management, with the potential to save lives and transform healthcare delivery.
Desiging secure campus network and mitigating network attacks
(NHCE, 2025) Janak Sapkota 1NH21CS107; Dipal Thapa 1NH21CS076; Pruthviraj K G 1NH21EC122; Rohit Shah 1NH21EC131
The project, "Designing Secure Campus Network and Mitigating Network Attacks," focuses on creating a secure, scalable, and resilient network infrastructure for educational institutions. It addresses challenges posed by increasing cyber threats targeting sensitive data and critical operations. The proposed solution incorporates VLAN segmentation, RADIUS authentication, and multi-factor authentication to enhance data security and restrict unauthorized access. Advanced Layer 2 protections, such as dynamic ARP inspection and DHCP snooping, safeguard against common vulnerabilities like ARP poisoning and rogue DHCP servers. Failover protocols, including HSRP and redundant firewalls, ensure high availability and operational continuity. Real-time threat detection, supported by AI-based monitoring and automated alerts, improves incident response, while future-ready technologies like SASE and machine learning enhance scalability and adaptability. Extensive simulations in virtualized environments validated the system, achieving 99.98% uptime, significant reductions in attack success rates, and enhanced overall performance. This project establishes a robust model for securing educational networks against evolving cybersecurity challenges.
Design and Development of autonomous vehicle for in campus Logistics
(NHCE, 2025) Adith sanjay EN 1NH21EC010 Aseel Javeed 1NH21EC024 Mohammed Huzaifa 1NH22ME408 Mohammad Yousuf 1NH22ME407
Autonomous vehicles are revolutionizing logistics operations across various sectors, including campus environments. This study focuses on designing and implementing an autonomous vehicle system tailored for in-campus logistics, operating in a closed environment with predefined routes. The proposed vehicle incorporates barcode scanners, cameras, and intelligent algorithms to navigate efficiently and safely, ensuring seamless delivery operations. The vehicle relies on barcodes strategically placed along its predefined route to obtain positional and directional information. The barcode scanner reads these markers, allowing the onboard algorithm to interpret and determine the next navigational step. This eliminates the need for complex GPS systems and ensures high precision, even in environments with limited satellite connectivity. Complementing this, the integrated camera system provides real-time obstacle detection and path verification. By fusing data from both systems, the vehicle achieves robust and dynamic navigation, maintaining its reliability even in unpredictable scenarios. This autonomous delivery system is designed to cater to various logistics tasks within campuses, such as delivering parcels, documents, or food. It prioritizes safety, efficiency, and scalability, enabling operations in diverse campus settings with minimal human intervention. The system is energy-efficient, leveraging electric power to align with sustainability goals while reducing operational costs. The project demonstrates the potential of combining advanced technologies, such as barcode scanning, computer vision, and machine learning, to create a cost-effective and user-friendly logistics solution. Future advancements, including AI-powered decision-making and multi-vehicle fleet management, could further enhance the system’s capabilities. This study sets a foundation for the broader adoption of autonomous delivery vehicles in structured environments, paving the way for innovative logistics solutions in educational, corporate, and institutional settings.
Hand-gesture controlled presentation
(NHCE, 2025) Achyuth M 1NH21CS005; Manjunath Naik 1NH21CS154; Sagargouda patil 1NH21EC134; Sujay S Bagalur 1NH21EC156
In the age of technological innovation, human-computer interaction has progressed towards touchless and intuitive interfaces. The Hand Gesture Controlled Presentation System is an innovative project that allows users to control presentation software using hand gestures. By integrating computer vision and machine learning techniques, this system eliminates the need for physical devices such as keyboards or mice, enabling seamless and contactless interaction. It is particularly valuable in scenarios where traditional input devices are impractical, such as during public presentations or in health-sensitive environments. The system utilizes real-time image processing to detect and interpret hand gestures, translating them into commands for controlling presentation slides. It employs Python as the programming language, using libraries like OpenCV, MediaPipe, and CVZone for efficient and accurate hand gesture recognition. These libraries enable real-time tracking of hand movements and gesture classification, ensuring precise and responsive interaction with the presentation software. The core functionality begins with hand detection, where the system captures video frames from a webcam and identifies the user's hand using advanced algorithms. MediaPipe, known for its robust hand-tracking capabilities, detects hand landmarks and tracks their movement in real-time. Once the hand is detected, the system moves to gesture recognition, identifying specific gestures based on the spatial arrangement of landmarks. These gestures are mapped to predefined actions, such as swiping left or right to navigate slides, pinching to zoom in or out, and waving to exit the presentation. One of the system's key strengths is its ability to operate in dynamic environments. Real-time processing ensures minimal latency, while the combination of OpenCV's image processing and MediaPipe's tracking algorithms enhances accuracy. CVZone simplifies the implementation of gesture-based interactions, allowing developers to focus on user experience rather than technical complexities.
Women safty night patrolling robot
(NHCE, 2025) B Sushma 1NH21EC027 S mahitha 1NH21EC139
Our project is based on Embedded Systems domain, Embedded systems are specialized computing systems designed to perform dedicated functions within larger mechanical or electronic systems. They play a vital role in robotics, offering control, automation, real-time processing, and sensor integration. These systems are compact, efficient, and reliable, making them ideal for applications like a women safety night patrolling robot. This robot is an innovative solution to enhance safety, particularly in high-risk areas, by combining hardware components such as microcontrollers, sensors, cameras, actuators, and communication modules with embedded software, AI algorithms, and protocols. The robot is equipped with motion sensors, cameras, audio sensors, to monitor its environment, detect potential threats, and respond accordingly. It uses autonomous navigation, powered by GPS and obstacle detection, to patrol predefined routes or dynamically adjust paths based on real-time inputs. Its surveillance capabilities are augmented by AI, which detects suspicious behavior or anomalies in video and audio data. In emergencies, the robot sends automatic alerts to authorities or predefined contacts, provide live video feeds, and includes audio and visual alarms for immediate response. Key functionalities include real-time monitoring, alert systems, two-way communication, and emergency response, all aimed at enhancing safety and situational awareness. Embedded systems enable these capabilities by seamlessly integrating hardware and software to ensure reliable operation. Despite challenges such as energy efficiency, data privacy, cost, and environmental adaptability, the robot offers numerous benefits, including scalability, deterrence of threats, and empowering individuals by fostering a safer environment.
Revolutionizing skull classification: leveraging digital forensics and deep learning in physical anthropology
(NHCE, 2025) C G Prasanth Reddy 1NH21EC036 C Anjan Reddy 1NH21EC039 Vasanth Kumar S 1NH21AI116 Abhilash Reddy B 1NH21AI146
This study explores the integration of digital forensics and deep learning to revolutionize skull classification within physical anthropology. It employs advanced image processing techniques and convolutional neural networks (CNNs) for accurate identification and classification of skulls, particularly based on the presence or absence of the mandible. The research process begins with preprocessing skull images, including resizing, noise removal, and contrast enhancement, to ensure high-quality inputs. Feature extraction techniques such as Gray-Level Co-occurrence Matrix (GLCM), Discrete Wavelet Transform (DWT), Gabor filters, and Segmentation-based Fractal Texture Analysis (SFTA) are used to capture intricate textural and structural details of the skulls. The study compares traditional methods utilizing Support Vector Machines (SVMs) with the proposed CNN-based approach, demonstrating the latter's superior performance with significantly enhanced accuracy metrics. Data augmentation techniques like rotation, flipping, and scaling further bolster model robustness and mitigate overfitting. Evaluation metrics, including accuracy, precision, recall, and F1 score, validate the methodology's effectiveness, achieving classification accuracies exceeding 99%. This innovative framework supports anthropologists, forensic scientists, and archaeologists in efficient management, analysis, and preservation of anthropological collections. Its application extends to museum collections, forensic investigations, archaeological research, and medical studies. The findings emphasize the transformative potential of integrating machine learning and digital forensics in anthropological research, offering a robust tool for advancing our understanding of human evolution and cultural heritage.
NEURONET: Convergence of brain-compuer interfacing and networking technology
(NHCE, 2025) Nandini R 1NH21EC103; Monisha M 1NH21EC098; Abdul Khadar 1NH21EE003; Yashas 1NH21EE126
The NeuroNet project explores the seamless integration of Brain-Computer Interfaces (BCI) with networking systems to perform real-world software operations. By simulating EEG signals and classifying them based on frequency ranges, the system leverages a Python-based client-server model to execute tasks. These tasks demonstrate the potential of BCIs in enabling non-invasive and efficient control of digital environments. The project emphasizes the practical application of networking in processing brain signals, paving the way for enhanced automation, accessibility, and control across domains such as assistive technology and smart systems. Through the NeuroNet framework, we aim to highlight the transformative potential of brainwave-driven digital solutions. Network protocols: Exploring suitable protocols for reliable and efficient transmission of brain signal data and control commands over a network, considering factors like latency, bandwidth, and security. Client-server communication: Designing robust and scalable communication mechanisms between clients and the server, ensuring seamless data exchange and synchronization. Application development: Developing a diverse range of applications demonstrating the potential of the system, such as remote control of assistive devices, collaborative virtual environments, and neurogaming. Usability and evaluation: Conducting user studies to assess the system's usability, performance, and overall user experience, identifying areas for improvement and optimization.
Diabetic retinopathy using deep learning - integration of CNN and SVM models
(NHCE, 2025) Abhiroop HR 1NH21EC005; Deepthi L 1NH21EC043; Shoaib Javeed 1NH21A1095, Yamini K 1NH21AI121
Diabetic Retinopathy (DR) is a progressive ocular condition resulting from diabetes, characterized by damage to the retinal blood vessels, leading to vision loss if not detected early. As DR progresses asymptomatically in the initial stages, early screening and accurate diagnosis are critical to prevent irreversible damage. Traditional diagnostic methods rely on manual examination of retinal images, which is time-consuming and prone to errors due to the subjective nature of interpretation. In response to these challenges, the integration of deep learning models, particularly Convolutional Neural Networks (CNNs) and Support Vector Machines (SVMs), presents a promising solution for automated DR detection. This study introduces a hybrid deep learning approach that combines CNNs and SVMs for DR diagnosis. The CNN model serves as a feature extractor, automatically learning relevant patterns and structures from retinal images. Its ability to process raw pixel data and hierarchically capture visual features makes it highly effective for detecting DR-related abnormalities such as microaneurysms, hemorrhages, and exudates. However, CNNs, although powerful, can sometimes suffer from overfitting and limited generalization to unseen data. To address these issues, we integrate the SVM classifier, known for its robustness in handling small, high-dimensional datasets and creating complex decision boundaries. The hybrid CNN-SVM model works in two stages: first, the CNN learns spatial features from the retinal images, and these features are then passed to the SVM, which performs classification by finding the optimal hyperplane in a transformed feature space. This approach effectively enhances the model's ability to discriminate between different stages of DR and reduces false positives and negatives. The proposed system is evaluated on benchmark DR datasets such as the EyePACS and APTOS datasets, which include images labeled for various stages of DR, from no DR to proliferative DR. Performance metrics such as accuracy, sensitivity, specificity, and F1-score are compared against baseline models and other state-of-the-art techniques. The results show that the CNN-SVM hybrid model outperforms standalone CNN models and traditional machine learning methods, achieving high accuracy and robust classification performance even in the presence of imbalanced data. In addition to providing a powerful diagnostic tool, this research underscores the potential of combining deep learning with classical machine learning techniques to enhance the reliability of medical image analysis. The proposed model offers a scalable, cost-effective solution for automated DR detection, enabling early intervention and reducing the burden on healthcare systems. Key words: Diabetic Retinopathy, deep learning, Convolutional Neural Networks, Support Vector Machines, hybrid model, retinal fundus images, automated diagnosis, feature extraction, medical image analysis, early detection, classification accuracy, microaneurysms, hemorrhages, exudates, healthcare AI, machine learning in healthcare, EyePACS dataset.

Browse

Browsing 2024-25 by Issue Date

Results Per Page

Sort Options