ONLINE CLASSIFIED CITY SEARCH ADS WEBSITE

Kalaiarasi P1 G.Gopi2 V.Jahnavi Sai Sree3 L.Amulya4 B.Ajay5

1,2,3,4,5 Department of Computer Science and Engineering, Kalasalingam Academy of Research and
Institution Anand Nagar, Krishnankoil, Tamilnadu, India

p.kalaiarasi@klu.ac.in, Gopigonuguntla197@gmail.com, Jahnavisaisree2002@gmail.com

Lakkuamulyareddy2@gmail.com , Bheemavarapuajay2018@gmail.com

Abstract
In today's digital landscape, online classifieds have emerged as a crucial solution, offering the
convenience of seamlessly transitioning between buyer and seller roles with just a click. Moreover,
being accessible globally, these platforms transcend geographical boundaries, enabling anyone,
anywhere, to peruse listings. Basic computer literacy is all it takes to navigate these sites. Introducing
the "Online Classified City Search of Products" project, which endeavours to establish a highly
efficient and user-centric online hub connecting individual sellers, shops, buyers, and administrators
within a specific city. Powered by Java Server Pages (JSP), this web-based application promises an
intuitive interface, facilitating seamless transactions. Tailored to accommodate diverse user roles,
including individual sellers, shops, buyers, and administrators, each with distinct privileges and
functionalities, the system empowers sellers and shops to register and present their products through
detailed listings encompassing descriptions, images, pricing, and contact details. Buyers, on the other
hand, enjoy the liberty to explore a wide spectrum of products within their city, refining their searches
with customizable filters based on categories and prices, ensuring a personalized shopping journey.
Meanwhile, administrative oversight guarantees platform integrity, encompassing user registrations
and upholding overall functionality and security standards.
 REAL-TIME CRYPTOCURRENCY TRACKER
Priyanka Behki1 Siddhant Aryan2 Raunak Raj3 Saurabh Kr. Singh4
Khushi Yadav5 Mohd. Suhail Saifi6

1,2,3,4,5,6 BE-CSE, Chandigarh University, Mohali,

behkipriyanka1998@gmail.com, siddhantaryan09@gmail.com, raunakraj650@gmail.com

saurabhsingh20pr@gmail.com, khushisai01@gmail.com, msuhailsaifi143@gmail.com

Abstract
This ReactJS-based web app provides live cryptocurrency metrics, including market capitalization
and 24-hour price changes. It also converts this data into visual graphs and ranks cryptocurrencies
based on their value. "Crypto Gyan" is a versatile platform designed to cater to the diverse needs of
cryptocurrency enthusiasts. Its feature-rich interface and user-friendly design make it a
comprehensive hub for up-to-date information, insights, and educational resources on Bitcoin and
other cryptocurrencies. To stay informed in the dynamic cryptocurrency landscape, users can explore
diverse topics including market dynamics, legal updates, and technological innovations. "Crypto
Gyan" offers comprehensive analytics that let users track live market prices, observe price
fluctuations, and compare trading volumes between currencies. By customizing alerts and
notifications, users can stay ahead of market trends and make informed investment decisions "Crypto
Gyan" empowers users with a comprehensive platform that caters to all skill levels in the dynamic
realm of digital assets. It offers a comprehensive learning experience, enabling beginners to grasp the
basics of cryptocurrencies and advanced traders to access insightful market analysis.
 Implementation and Analysis of Human Emotion Using deep learning
techniques

Aakash Kumar1 Mukund Kumar Mishra2 Mr. Damodharan D3

Dept. of Computing Science & Engineering, Galgotias University, Greater Noida, Uttar Pradesh,
India,
aakash555kumar@gmail.com, mukundmishra815@gmail.com,
damodharan@galgotiasuniversity.edu.in

Abstract
Face expressions frequently convey shared emotional states, and this similarity has led to the
widespread usage of face recognition technology in a variety of security and personal information
applications. By using facial expression recognition, it is possible to identify emotions such as fear,
rage, surprise, sorrow, and happiness. This provides insightful information about an individual's
emotional state. Recognition and detection of facial emotions are important in the field of marketing,
as client feedback on a business's goods and services is essential to its survival. Artificial intelligence
systems may be used to analyse the emotional reactions recorded in photos or videos, making it
possible to measure customer happiness. Previously, modified photos were used to evaluate emotions
using machine learning techniques like random forest and support vector machines (SVM). Deep
learning's latest breakthroughs have contributed to the area's developments in computer vision. Facial
expression detection using Convolutional Neural Network (CNN) models has shown efficacy; these
models require a dataset for training and testing.
 A Systems-Aware Approach to Data Infiltration Attacks on collaborative
Machine Learning Approach

Koppisetti Venkata mani Madhuri1 Dr P R Sudha Rani2

1
PG Scholar, Computer Science and Engineering, Shri Vishnu Engineering College for women,
Bhimavaram, India ,
2
Professor, Computer Science and Engineering, Shri Vishnu Engineering College for women,
Bhimavaram, India

madhurikoppisetti2000@gmail.com prsudharanicse@svecw.edu.in

ABSTRACT
This study explores the vulnerability of collaborative multi-task learning, where various devices
cooperate to train a shared model while preserving user data privacy, to data manipulation attacks.
Attackers exploit communication channels to inject corrupted data, undermining the integrity of deep
learning models. The research formulates effective data poisoning attacks as a two-part optimization
program, introducing a new optimization method called Attack to Collaborative Learning (A2CL).
A2CL efficiently computes hidden influences for poisoned data, facilitating effective attack strategies
in collaborative learning. Demonstrations reveal high susceptibility to poisoning attacks, emphasizing
the urgent need for robust security measures in collaborative learning systems. This study marks a
pioneering investigation into data poisoning attacks in this collaborative learning paradigm, paving
the way for further research into secure collaborative learning frameworks.
DETECTING DIGITAL FINANCIAL TRANSACTION FRAUDS USING
CNN MODEL

Chadalavada Ester Rani1 V. Purushothama Raju2

1
PG Scholar, Dept. of Computer Science & Engineering, Shri Vishnu Engineering College for
Women(A), Bhimavaram, A.P, India
2
Professor, Dept. of Computer Science & Engineering, Shri Vishnu Engineering College for
Women(A), Bhimavaram, A.P, India

erani8982@gmail.com praju@svecw.edu.in

Abstract
The Unified Payments Interface (Financial Transaction) has become increasingly popular as a
convenient and user-friendly digital payment platform. However, the surge in Digital Transaction has
also resulted in a rise in fraudulent activities, presenting a significant security challenge. This paper
introduces a novel approach to detecting fraudulent Digital Transaction by leveraging convolutional
neural networks (CNN). By incorporating the CNN model, this study aims to offer a cost-effective
solution for identifying Digital Financial fraud. The CNN's ability to analyse patterns in data makes it
particularly suitable for fraud detection in Digital Financial Transactions. We evaluate our proposed
methodology using a comprehensive dataset that includes both fraudulent and genuine transactions.
Our results demonstrate a significant improvement in fraud detection accuracy and efficiency
compared to existing methods. This research contributes to the advancement of secure digital payment
systems by presenting an innovative approach to detect fraud transactions. Financial institutions and
payment service providers can utilize this methodology to enhance the security of Digital Financial
Transactions and protect users from fraudulent activities.
 Steganography and Data Encryption in Images Using Contrast Stretching
and AES Encryption
Puneet Bhardwaj1, Sunita Singhal2, Deepika Shekhwat3

1,2,3 Department of Computer Science and Engineering, Manipal University Jaipur, Jaipur, India,
Amity School of Engineering and Technology, Mumbai, Maharashtra

puneet.219301298@muj.manipal.edu, sunita.singhal@jaipur.mainpal.edu,
dshekhawat@mum.amity.edu

Abstract
Steganography is an important technique in data security, and its purpose is to hide text, especially
images, in another environment. This paper presents a new method that uses variable stretching to
identify low areas in images to ensure low noise during data acquisition. Private measurements are
encrypted using Advanced Encryption Standard (AES) and stored in image metadata, ensuring the
security and integrity of confidential information. The proposed method shows superior performance
in terms of efficiency, capacity (50% increase) and privacy compared to traditional methods such as
least significant bit (LSB) steganography. Our contribution includes the development of advanced
steganography algorithms that place information in low-distortion areas, making them more robust
against detection and attack. We also provide a comprehensive analysis and comparison with existing
methods showing the progress of our security and data integrity methods. These improvements
address the limitations of current technology and provide a safer and more reliable way to hide
information in digital images.
 Integrating Emotional Analysis with Advanced Driver’s Drowsiness
Detection System using Convolutional Neural Networks

Mandapati Ankitha1, Sandeep Vemuri2, Jarabala Sindhu Bhargavi3 and Suraneni Lowkya Gayathri4
1,2,3,4 Velagapudi Ramakrishna Siddhartha Engineering College, Vijayawada, India

ankithamandapati2003@gmail.com, sandeep.vemuri@gmail.com,
jarabalasindhubhargavi@gmail.com , lowkyagayathri@gmail.com

Abstract
Traditional drowsiness detection systems often lack emotional context, which can significantly impact
driver safety. To address this, the proposed system integrates emotional analysis, enabling a clearer
way of understanding the driver's state. The initial stage employs the Haar Cascade model to detect
the driver's face, facilitating subsequent tracking and analysis. Lukas-Kanade Flow and ShiTomasi
corner detection are applied to monitor head movements, enabling the identification of signs of
drowsiness. This dynamic input is coupled with a combination of CNN networks to effectively
capture spatial and temporal features from facial images and sequences. This fusion enhances the
system's accuracy in identifying drowsiness-related patterns. To validate the effectiveness of the
proposed approach, extensive experiments are conducted with a diverse dataset Cohn-Kanade dataset
for emotional analysis and Drowsiness Detection Dataset (DDD). The results obtained from the CNN
performance metrics are evaluated against existing drowsiness detection methods. Through these
evaluations, the study aims to determine the optimal algorithm for accurately detecting and predicting
driver drowsiness, thus contributing to the development of advanced driver assistance systems
(ADAS) and improved road safety measures.
 Building an Anonymous Feedback website with MERN

Akash Shukla1 Dr. Ashish Tiwari2

1,2 Department of Computer Science and Engineering, Amity University, Lucknow,

akashshukla2311@gmail.com atiwari3@lko.amity.edu

Abstract
In this research paper, we provide an in-depth analysis of creating an anonymous feedback website
using the Express.js, React.js, MongoDB, and Node.js MERN stack. The goal is to establish a forum
where people may offer input without disclosing their identity, encouraging open and sincere
dialogue. This project aims to improve user happiness and engagement by addressing the growing
requirement for privacy and secrecy in feedback systems. The MERN stack was selected as the best
option for this application due to its reliable design, scalability, and effective development process.
We offer a thorough examination of building an anonymous feedback website with the Express.js,
React.js, MongoDB, and Node.js MERN stack in this research paper. The idea is to provide a place
where people may contribute ideas without having to reveal who they are, promoting honest and
candid communication. The goal of this project is to meet the growing need for privacy and secrecy in
feedback systems in order to increase user satisfaction and engagement. The MERN stack's
dependable design, scalability, and efficient development process made it the greatest choice for this
application. User acceptability testing, performance testing, security testing, and other testing
approaches are used to assess the efficacy and functionality of the anonymous feedback website. The
outcomes show that the platform can manage a sizable amount of feedback while upholding user
privacy and data security. This study adds to the expanding body of information about anonymous
feedback systems and offers a workable option for companies looking to use cutting-edge web
development technology to deploy such systems.
 Energy-Efficient Clustering Algorithms for Wireless Sensor Networks: A
Comparative Study

Suresh Kumar Mandala1, Nagaraju Baydeti2 and Shakila Baskaran3

1,2 Department of Computer Science and Engineering, National Institute of Technology Nagaland,
Chumukedima, Dimapur 797103, Nagaland, India
3 Department of Electrical and Electronics Engineering, National Institute of Technology Nagaland,
Chumukedima, Dimapur 797103, Nagaland, India,

mandala.suresh83@gmail.com , baydetinagaraju@nitnagaland.ac.in
baydetinagaraju@nitnagaland.ac.in

Abstract:
Wireless Sensor networks, or WSNs, are now an important part of many things, from keeping an eye
on the military to farming more accurately and protecting the environment. Sadly, the batteries in
these sensor nodes don't last very long. It is crucial to conserve power since it impacts the network’s
performance and durability. Here are five well-known WSN clustering algorithms that will be
examined. Some of them are LEACH (Low-Energy Adaptive Clustering Hierarchy), DEEC
(Distributed Energy-Efficient Clustering), TEEN (Threshold-sensitive Energy Efficient Network),
HEED (Hybrid Energy-Efficiency Distributed Clustering), and Fuzzy Logic-based Clustering. A
framework for network simulations that was made to look like a typical WSN work environment was
used for the simulations. TEEN was found to be the best algorithm in terms of how long the network
lasts and how much power it runs on. In different situations, fuzzy logic-based clustering, TEEN, and
LEACH have been found to be the best ways to make networks last longer.
 Aerial Data Protection Using Blockchain for Drone Communication
Mr. Chiranjit Das1, Dr. Arti Patle2, Pratik Bujade3∗, Anand Kadav4, and Aditi Mali5
1
G H Raisoni College of Engineering and Management, Pune, India
chiranjit.das@ghrcem.raisoni.net 2 drartipatleyede@gmail.com
3
pratikbbujade@gmail.com
4
anandkadav14@gmail.com
5
aditimali1704@gmail.com
ghrcemp.raisoni.net

Abstract
Drones are transforming many industries with its rapid progress and wide range of application, but
their communication systems remain vulnerable at many points. Current systems are centralized
which are prone to cyber-attacks, security threats, data manipulation, and single points of failure due
to their centralized nature. To address this encounter, our project offers a decentralized and
autonomous drone communication system that leverages the blockchain technology and drone
simulation to ensure delivering the secure, efficient, and distributed communication among drones. By
providing a tamper-proof record of drone communication, our system has wide range of implications
for various industries, including surveillance and intelligence gathering, rescue and disaster response,
precision agriculture, wildlife conservation, research and information gathering, infrastructure
inspection, and environmental monitoring. Our state-of-the-art approach enables real-time situational
awareness, informing strategic decision-making and improvised operational efficiency for
organizations and industries where quick and errorfree response are required.
 ProSkillz - An online Service Provider Application
Deep Kawale1 , Ninad Gaonkar1 , Junaid Ansari1 , and Prof. Sumitra Sadhukhan1
1
Department of Computer Engineering, Rajiv Gandhi Institute of Technology, Mumbai, India
kawaledeep9@gmail.com, ninadgaonkar234@gmail.com, junaidshahid.ansari@gmail.com,
sumitrassadhukhan@gmail.com

Abstract
This study introduces ”ProSkillz,” a mobile application designed to connect users with local service
providers for home-based services. ProSkillz enables users to schedule any possible services such as
cleaning, laundry, and electronics repair, similar to other platforms. However, its unique selling point
is the promotion of all local service providers, giving neighborhood businesses greater visibility. The
app utilizes Node.js for backend operations, Express.js for server-side scripting, GraphQL for efficient
data queries, React and React Native for user interfaces, and MongoDB for data storage. It also uses
the Haversine formula for precise distance calculations, allowing users to search for services and view
providers on a map. With an easy-to-use interface, users can book appointments, while providers can
list their services and locations. ProSkillz offers a community-focused, efficient platform that
contributes to the growing trend of on-demand service applications.
 Staying Awake Behind the Wheel: A Systematic Review of Monitoring
Techniques for Drowsiness
Sumanta Chatterjee1, Amit Majumder2
1
Research Scholar, Dept of CSE, NIT Jamshedpur, Jharkhand, India.
2
Asst.Professor, Dept of CSE, NIT Jamshedpur, Jharkhand, India
sumanta.prof@gmail.com

Abstract
Driver drowsiness has increasingly become a significant contributor to road accidents, resulting in
severe injuries, fatalities, and substantial economic repercussions. This paper addresses the urgent
need for efficient detection techniques that can identify driver fatigue in real-time, potentially
preventing many accidents linked to drowsiness. We conduct a comprehensive survey of various
designs and methods for detecting drowsiness, examining both traditional approaches—such as
physiological indicators (e.g., eye blink rate, heart rate variability) and behavioral metrics (e.g.,
steering wheel movement, lane deviation)—as well as modern solutions utilizing advanced
technologies like machine learning and computer vision. Our comparative analysis evaluates the
effectiveness, limitations, and integration potential of these techniques within existing vehicle safety
systems. The findings underscore the importance of timely = drowsiness detection in enhancing road
safety and reducing the incidence of fatigue-related accidents.
 LDA based Face Recognition Attendance Management System
M. Shyamala Devi, Department of Computer Science and Engineering, Panimalar Engineering
College,Chennai, Tamilnadu, India, shyamalapmr@gmail.com
R.Sarmilaa, Department of Computer Science and Engineering, Panimalar Engineering College,
Chennai, Tamilnadu, India, samkaviravi23@gmail.com
N.Sai Sruthi, Department of Computer Science and Engineering, Panimalar Engineering College,
Chennai, Tamilnadu, India, sruthipriya6639@gmail.com

Abstract
Facial recognition, a secure biometric technique, typically requires large databases and is often
outsourced by large companies. Effective attendance management systems have gradually become
progressively more important in organisations and schools in recent years. This research offers a novel
face recognition attendance management system based on Linear Discriminant Analysis (LDA). The
suggested solution streamlines the attendance monitoring procedure by integrating cutting-edge image
processing algorithms to reliably capture and recognise people in real-time. This work aims to
develop an in-house system for educational institutions, allowing student attendance to be recorded
through facial recognition. Each student’s facial data is stored as 100 different images, with
attendance records logged with precise timestamps. The first step of the system is to gather participant
face pictures, which are then processed to improve quality and lower noise. Then extract the most
discriminative characteristics from the facial data using LDA, allowing for reliable identification in a
variety of illumination and angle conditions. Following processing, the images are compared to a
database that has been maintained, allowing for instant attendance tracking and identification. The
initial outcomes showcase the efficacy of the LDA-based methodology, attaining elevated levels of
98% accuracy in facial identification while drastically decreasing the duration needed for attendance
documentation in contrast to conventional techniques. This solution improves security and
accountability while reducing human error. Institutions may concentrate more on operational
effectiveness and educational outcomes by automating the attendance process. Further investigation
attempts will examine the incorporation of other biometric modalities and the implementation of the
system in heterogeneous settings.