Project Reports - Computer Engineering
Permanent URI for this collection
Browse
Recent Submissions
Item AI-based multimodal deepfake detection and prevention system(AIKTC, 2026-06) Ansari, Mukhtar; Khan, Kaif [22CO19]; Mohd. Aafaque [22CO31]; Shedge, Omkar [22CO34]; Khargaonkar, Sandesh [22CO40]The blistering development of artificial intelligence and deep learning has provided the possibility to create most convincing synthetic media, which is often known as deepfakes. There are increasingly greater threats to privacy, security and digital authenticity that such manipulated images, videos and audio clips present. This project presents an AI-Based Multimodal Deepfake Detection and Prevention System that is effective in detecting and eliminating falsified media through machine learning and neural network models. The proposed system combines EfficientNet and ResNet systems to authenticate images and videos and use recurrent neural networks (RNNs) and spectrogram-based analysis to identify modified audio. These modules are integrated into a web interface running on Flask, by which users could upload multimedia contents and automatically check their authenticity. The system enhances detection accuracy in the various modalities through pre-processing, feature extraction, and classification. Such preventative methods as digital watermarking and metadata embedding are also discussed in order to preserve the reliability of the media. Altogether, this piece is a step towards building credible digital ecosystems and combating misinformation that can result due to AI-generated content. The project improves the accuracy of the detection of several modalities by preprocessing, feature extraction, and deep learning classification. Also, such preventive steps as watermarking and embedding secure metadata are discussed to decrease the abuse of AI-generated content. This work helps to empower the integrity of digital media, ethical use of AI, and protect individuals and organizations against misinformation and identity distortion. Key Features : 1. Multi-Mode Detection- Detection of deepfakes in images, videos, etc. and audio content.Item Sci tech web crawler(AIKTC, 2026-06) Ansari, Mukhtar; Wangde, Amin [21CO62]; Shaikh, Ashraf [22CO44]; Mohammed Ali Sahabuddin [22CO50]; Khan, Mohd. Afzal [20CO30]The Sci-Tech Web Crawler project is designed to build a domain-specific search engine focused exclusively on science and technology content. Unlike general-purpose crawlers that index a wide range of web pages based on popularity or backlinks, this system prioritizes relevance, credibility, and academic quality. It gathers and analyzes information from authoritative platforms such as IEEE, arXiv, and other verified research repositories to ensure that users—students, researchers, and professionals—gain access to accurate and topic-specific data. The primary objective of this work is to create a precise, intelligent, and automated tool that minimizes irrelevant or promotional noise typically present in mainstream search engines. To achieve this goal, the crawler leverages Python-based technologies including BeautifulSoup, Requests, Scikit-learn (TF-IDF), NLTK, and Hugging Face Transformers. The methodology involves fetching HTML content from user-specified URLs, extracting and cleaning text, computing TF-IDF scores to identify high-value keywords, and generating concise summaries using stateof- the-art language models. The processed information is stored efficiently in MongoDB, forming a structured knowledge base that can be queried and expanded dynamically. Additionally, a separate module categorizes hyperlinks by thematic relevance—such as Artificial Intelligence, Machine Learning, or Quantum Computing—enabling a more organized and domain-aware exploration of scientific resources. Through this implementation, the project successfully demonstrates how artificial intelligence and natural language processing techniques can be combined to build an intelligent, research-oriented web crawler. The system not only achieves its objective of focused data extraction but also offers scalability for future integration with real-time academic databases and interactive search interfaces. Overall, Sci-Tech Web Crawler contributes meaningfully toward improving the accessibility, accuracy, and efficiency of scientific information retrieval in the modern digital era.Item Machine learning- powered intelligent CCTV for crime detection and proactive security(AIKTC, 2026-06) Jamkhandikar, Irfan; Bhosale, Mrunal [22CO02]; Baditya, Deepak [22CO13]; Gaikwad, Sankalp [22CO17]; Madar, Adnan [22CO25]Traditional CCTV systems depend on continuous human monitoring, often leading to delayed threat recognition and ineffective responses. This project presents a Smart CCTV Application Extention that transforms conventional surveillance into an intelligent, real-time monitoring system using artificial intelligence and machine learning. The system integrates advanced computer vision and deep learning models YOLO, CNN, and Lip- Net to automatically detect threats such as weapon presence, unauthorized intrusion, and shoplifting, while also interpreting lip movements to identify potentially harmful verbal cues. Once suspicious activity is detected, automated alerts are generated to notify relevant authorities for immediate action. Designed for compatibility with existing CCTV infrastructure, the system offers a cost-effective and scalable upgrade requiring no hardware modifications. Experimental evaluations demonstrate high detection accuracy, efficient real-time processing, and adaptability across diverse environments. This approach enhances situational awareness, minimizes human dependency, and represents a significant step toward autonomous and proactive security systems suitable for both public and private surveillance applications.Item AI powered depression detection(AIKTC, 2026-06) Sadruddin, Safia; Shaikh, Muhammad Ismail (21CO54); Maruf Soban Wajuddin (21CO58); Kapadia, Mohammed Anas (22DCO01); Patel, Muskan (22DCO04)In today’s digital era, school certificates such as SSC marksheets are widely used for admissions and job applications, but verifying their authenticity remains a challenge. Traditional verification methods are often manual, time-consuming, and prone to forgery. To address this issue, this project presents SmartCert: AI and Blockchain-Based Certificate Verification System, focused on secure verification of SSC certificates. The system uses Blockchain technology to store certificate data in a tamper-proof and decentralized manner, ensuring that once the data is recorded, it cannot be altered. Additionally, Artificial Intelligence is used to identify irregular patterns and help detect possible fake certificates. A simple and user-friendly interface allows authorities to upload SSC certificate data and users to verify it quickly and securely. Overall, the system provides a faster, more reliable, and secure method for SSC certificate verification, reducing manual effort and minimizing the risk of fraudItem Electricity theft detection(AIKTC, 2026-06) Kazmi, Kashif; Kazi, Omeir Asim [22CO33]; Shaikh, Faizan [22CO46]; Shaikh, Qizarhayat [22CO56]; Shinde, Abhijeet [22CO58]This project addresses the critical challenge of electricity theft in power distribution systems, which causes significant financial loss and grid instability. Given the scarcity of real-world labeled fraud data, the primary objective was to develop a robust detection system capable of identifying illicit consumption patterns using smart meter data. Methodology The objective was achieved through a multi-staged technical pipeline: Data Processing : Raw smart meter data (collected at 3-minute intervals) was aggregated into daily consumption windows. To overcome the lack of labeled data, we designed and injected realistic theft scenarios— including intermittent reduction, sustained drops, and partial manipulation— into the clean dataset. Feature Engineering: We extracted a comprehensive suite of behavioral features, such as day/night consumption ratios, rolling statistical variances, and ”low-usage streaks,” to capture deviations from normal consumer habits. Hybrid Ensemble Modeling: We implemented a two-layered classification architecture. A Supervised Layer (Random Forest and XGBoost) identifies known theft signatures, while an Unsupervised Anomaly Layer (Isolation Forest and Local Outlier Factor) detects novel or statistical outliers. Risk Categorization: Results are merged into a three-tier risk model: Normal, Suspicious, and High Risk.Item Securing the openplc(AIKTC, 2026-06) Kazi, Samreen Banu; Bronkar, Basheer [23DCO01]; Sayyed, Mohammad Kaif [22CO42]; Shaikh, Zaid [23DCO06]; Patel, Ahmed [22CO38]This project presents a multi-layered security enhancement for Industrial Control Systems (ICS) by hardening the OpenPLC platform against contemporary cyber-attacks. The primary objective is to address the inherent security deficits in legacy industrial protocols, such as Modbus/TCP and DNP3, which traditionally lack native authentication and encryption. To achieve this, we engineered an integrated security suite comprising Time- Based One-Time Password (TOTP) Multi-Factor Authentication for the web management interface, an automated vulnerability auditing tool for system boot-up, and TLS 1.3 encryption for data transit. The methodology further incorporates a real-time Intrusion Detection System (IDS) using Snort to perform Deep Packet Inspection (DPI), specifically targeting unauthorized industrial command injections. This objective was successfully achieved, resulting in a controller appliance capable of resisting sophisticated network-based threats. Experimental evaluations demonstrate that the proposed framework successfully identifies 98 percent of unauthorized industrial commands and completely mitigates credentialbased attacks. Significantly, the system maintains high operational efficiency by introducing a negligible latency overhead of less than 5 milliseconds, ensuring that the critical timing requirements of real-time industrial processes are never compromised.Item Smart ERP solution for academic eligibility assessment through machine learning(AIKTC, 2026-06) Nusrat Jahan; Bhoir, Priyanshi [22CO01]; Pathan, Mahenoor [22CO05]; Chavan, Aashish [22CO07]; Solkar, Aarhaan [22CO62]The growing volume of data and many educational institutions still relying on manual or rule-based methods pose major challenges in handling and assessing student academic eligibility. These methods can be timeconsuming, error-prone, and lack the ability to predict. This project is based on a Smart ERP Solution for Academic Eligibility Assessment using Machine Learning, which will help automate and improve the process of academic eligibility assessment by incorporating predictive analytics within an ERP system. The main goal of this research is to design a smart system that can process academic data of students and decide student eligibility in a rule-based as well as machine learning based approach. Our system adopts a hybrid approach, combining the institution’s eligibility rules with a Random Forest-based eligibility predictor aimed at enhancing decisionmaking and managing intricate cases. The process includes data gathering, cleaning, feature engineering, model training, hybrid decision-making, and visualizing the results in a user-friendly way for administrators and students. The proposed system was successfully deployed and has shown greater efficiency, lesser manual labour and better accuracy than existing systems. It also facilitates early detection of students in need of intervention. These findings demonstrate the integration of machine learning and ERP system enhances academic eligibility assessment. In summary, this project succeeds in its aim to convert conventional academic management systems into smart, automatic and data-driven systems, thus helping to improve the efficiency and transparency of academic processes.Item Classification of respiratory diseases using lung sounds(AIKTC, 2026-06) Jamkhandikar, Irfan; Chandiwala, Saniya [22CO03]; Mohammad Asim [22CO28]; Shaikh, Mohd. Anas [22CO52]; Mohammed Ali [22CO49]Item Real-time plant disease detection and ai-powered solution system(AIKTC, 2026-06) Qureshi, Rehaal; Khan, Ehtesham [22CO18]; Siddiqui, Amil [22CO59]; Karche, Shahid [23DCO02]; Parkar, Uzair [21CO33]This project focuses on developing a Plant Disease Detection System using Convolutional Neural Networks (CNN) which a type of deep learning model widely used in image classification tasks. The system can accurately identify and classify various plant diseases from leaf images. The model is trained on a large dataset containing images of healthy and diseased leaves of multiple crops such as apple, tomato, potato, corn, grape, and more. It can detect up to 30 different plant diseases with high accuracy. By using image processing and deep learning techniques, this model helps farmers and agriculturists quickly identify plant diseases and take preventive measures, thereby reducing crop loss and improving productivity. The trained CNN model is integrated into a simple web interface (or mobilefriendly app) where users can upload a leaf image to get an instant prediction of the disease along with its name and confidence score.Item Smart ambulance allocation(AIKTC, 2026-06) Khan, Tabrez; Panchkutti, Nabil [22CO35]; Mohamed Fazil [22CO27]; Khan, Mohd. Aaquib [22CO21]; Shaikh, Mohd. Zain [22CO53]This project presents a Smart Ambulance System designed to improve emergency medical response by reducing ambulance dispatch time and enhancing coordination among users, ambulances, and hospitals. The system allows users to report accidents through a chatbot interface by providing details such as location, accident description, and number of injured people, which are then sent to the backend for processing. A logical decisionmaking approach is used to evaluate the severity of the situation, based on which the system selects the nearest available ambulance and identifies the most suitable hospital considering factors like distance, availability, and real-time conditions. The system also calculates the Estimated Time of Arrival (ETA), including the time required to reach the accident spot and transport the patient to the hospital, while taking traffic conditions and route optimization into account. After processing all the information, the ambulance is automatically dispatched, and a live tracking link is generated and shared with the user, ambulance team, and hospital for real-time monitoring of location and arrival time. This system improves communication among all stakeholders, reduces response time, and enhances resource allocation, thereby providing a scalable and efficient solution for modern emergency healthcare management.Item DSD soft(AIKTC, 2023-05) Ansari, Abdul Majid; Pangarkar, Rabiya (19CO05); Khan, Shoab (19CO33); Khot, Shehzad (19CO35); Qurishe, Mohd Ziyaulhaq (19CO37)Direct store delivery (DSD) software is a powerful solution that streamlines the delivery process for businesses that deliver products directly to stores or customers. This software system architecture includes components such as mobile devices, inventory management, route optimization, payment processing, and reporting and analytics. By integrating these components, DSD software provides businesses with real-time visibility into their delivery operations, enabling them to optimize routes, manage inventory, and improve overall efficiency. With the ability to automate tasks, eliminate manual processes, and increase accuracy, DSD software can help businesses reduce costs, improve customer satisfaction, and gain a competitive edge in the marketplace.Item Didactic solutions(AIKTC, 2023-05) Qureshi, Rehaal; Khan, Muskan (19CO02); Khan, Sana (19CO03); Shaikh, Zikra (19CO08); Khan, Aftab (19CO23)Didactic Solutions As technological advancements continue to reshape the corporate landscape, it is crucial to update training methodologies to improve work efficiency. E-Learning has emerged as a vital component of employee training and internal human resource development. In recent years, web-based training has gained immense popularity in the corporate world as a cheaper, faster, and more efficient way to train large numbers of employees and individuals worldwide. Blended learning, a combination of face-to-face and online modalities, has also taken on a more important role in corporate training program design and implementation, providing a more flexible learning modality. We developed a service-based project, the Didactic Solutions, designed to provide training to both students and employees. The Didactic Solutions offers free trials and low-cost courses, leveraging user-friendly web-based training. The project platform offers live online training, and students can access recorded videos of their sessions upon successful completion of their courses. The project chatbot feature enables students to ask their instructor any questions they may have. In this paper, we provide an in-depth analysis of the design and implementation of the Didactic Solutions, and highlight the benefits of web-based training for corporate training programs. The project research findings suggest that the Didactic Solution’s design and implementation offer an efficient and cost-effective way to deliver high-quality training programs to students and employees, resulting in improved work efficiency and productivity.Item Eyedesk(AIKTC, 2023-05) Ansari, Abdul Majid; Shaikh, Nadir (15CO38); Sayyed, Sohail (18CO48); Kadav, Rutuja (19CO01); Thube, Vaishali (19CO09)Improving efficiency and performance is an important topic in the world today.Our project Eyedesk includes registration of patients, storing their details into the system, and also booking their appointments with doctors.Our state-of-the-art facility is equipped with the latest technology, and our highly skilled and experienced team of ophthalmologists, optometrists, and support staff are dedicated to ensuring that you receive the best possible care.Our software has the facility to give a unique id for every patient and stores the details of every patient and the staff automatically. User can search availability of a doctor and the details of a patient using the id. Eyedesk can be entered using a username and password. It is accessible either by an administrator or receptionist. Only they can add data into the database. The data can be retrieved easily. The interface is very user-friendly. The data are well protected for personal use and makes the data processing very fast.It is having mainly two modules. One is at Administration Level and other one is of user I.e. of patients and doctors. The Application maintains authentication in order to access the application. Administrator task includes managing doctors information, patient’s information. To achieve this aim a database was designed one for the patient and other for the doctors which the admin can access. The complaints which are given by user will be referred by authorities. The Patient modules include checking appointments, prescription. User can also pay doctor’s Fee online.At our Eye Hospital, we understand that your vision is precious, and we are committed to helping you maintain it.Item Voice controlled e-commerce(AIKTC, 2023-05) Nusrat Jahan; Mapari, Simran (19CO04); Shaikh, Alfiya (19CO62); Aga, Fiza (20DCO01); Patil, Aditi (20DCO07)This project aims to develop a voice-controlled e-commerce platform that enables users to shop for products using voice commands. The project utilizes modern natural language processing and voice recognition technologies to enable seamless interaction between users and the platform. The platform is designed to enhance the shopping experience by making it more convenient and accessible. The report outlines the development process, including the technology stack used, the system architecture, and the user interface design. Additionally, the report evaluates the system’s performance and user experience, identifying key challenges and future directions for platform improvement. The project highlights the potential of voice-controlled e-commerce platforms in revolutionizing the way people shop online and demonstrates the importance of incorporating cutting-edge technologies in creating an efficient and user-friendly shopping experience. The main objective of this project is to develop a web application where a voice assistant is integrated using Alan AI to present a voice-controlled commodity purchase as per the user’s request.Item Product recommendation system using ML(AIKTC, 2023-05) Dange, Anas; Baig, Sahil (19CO49); Sirkazi, Ruzbihan (19CO48); Mohd., Aasif (19CO36); Ulde, Kais (19CO46)Product Recommendation System Using ML In today’s modern epoch of information technology, the idea of efficiently finding one’s favourite product in a large dataset of application database, becomes an essential issue to address for the online content providers in order to attract the masses as opposed to their competitors. Recommender systems or recommendation systems, as they are popularly known, are information filtering systems which are usually integrated with several consumer and com- mercial applications. Such systems act as a bridge between various content facilitators such as social media websites, e-commerce portals, streaming platforms, etc. and the users of these applications, by suggesting them items from the application database which conform to the user preferences and past activities. Such personalized systems play a vital role, especially when the user is unclear of the item to the searched for. These systems are infiltrating every aspect of our lives, in the form of ?Because you watched? header on Netflix, ?People you may know? section on Facebook, ?Customers who bought this also bought’ partition on Amazon.Item Disease prediction using machine learning and image processing(AIKTC, 2023-05) Dange, Anas; Haddadi, Amir (19CO19); Khan, Shahrukh (19CO51); Husen, Mohd Altaf (19CO41); Qureshi, Mohd Zaki (19CO40)Current machine learning models for healthcare analysis focus on one disease per analysis, such as diabetes, liver, malaria, pneumonia diseases. Our project aims to predict multiple diseases using machine learning algorithms, streamlit, Flask API, and Python pickling to save and load the model's behavior. By analyzing all parameters that cause diseases, our system can detect the maximum effects a disease will cause. A study using a large medical image dataset trained a deep learning model to predict diseases using transfer learning techniques. The proposed approach achieved high accuracy, outperforming other traditional methods, and has potential applications in clinical settings to reduce human error, improve diagnostic accuracy, and reduce the time required for diagnosis. This project can help people by monitoring their condition and taking necessary precautions to increase life expectancy. Overall, this study demonstrates the effectiveness of using machine learning and image processing for disease prediction and provides valuable insights into future research in this area.Item E-commerce Dapp(AIKTC, 2023-05) Mishra, Richa; Kadiri, Saud (19CO21); Khan, Anas (19CO24); Khan, Farhaan (19CO26); Shikalgar, Farhan (19CO56)Improving efficiency and performance is an important topic in the world today. As it is well-known, cooperative computing is an effective and traditional approach, and it is widely used in various fields. Inspired by this idea, take E- commerce for example, Security is one of its important indicators. In E-commerce, the security technology has become a major issue restricting the rapid development and popularization of E- commerce.With the advent of new technologies, Blockchain plays a major role in ecommerce sector.With the characteristics of decentralization, persistency, anonymity and auditability, blockchain technology is a new tool to solve the product traceability, information security and privacy, payment efficiency and cost reduction in cross- border e-business.Existing E-commerce models are trapped in a dilemma between the proof of ownership and privacy protection. To address this issue,We have made a platform Shoppingverse which is a blockchain based ecommerce.we design a privacy-preserving business protocol by employing private smart contracts in the negotiation phase. The protocol allows counterparties make deals without the disclosure of private information such as identities, addresses, and phone numbers. Moreover, we employ the zero- knowledge proof to guarantee the ownership.Item Stock analysis and prediction website(AIKTC, 2023-05) Mishra, Richa; Shaikh, Muhammad Huzaifa (18CO40); Bagdadi, Ameeruddin (19CO12); Khan, Raquib (19CO31); Qureshi, Mohammed Usman (19CO47)In this project we attempt to implement machine learning approach to predict stock prices. Machine learning is effectively implemented in forecasting stock prices. The objective is to predict the stock prices in order to make more informed and accurate investment decisions. We propose a stock price prediction system that integrates mathematical functions, machine learning, and other external factors for the purpose of achieving better stock prediction accuracy and issuing profitable trades. There are two types of stocks. You may know of intraday trading by the commonly used term ”day trading.” Interday traders hold securities positions from at least one day to the next and often for several days to weeks or months. LSTMs are very powerful in sequence prediction problems because they’re able to store past information. This is important in our case because the previous price of a stock is crucial in predicting its future price. While predicting the actual price of a stock is an uphill climb, we can build a model that will predict whether the price will go up or down. Apart from that, we have included a stock analysis feature, which helps users to understand the performance of a particular stock, also it guides users to invest in a stock whose price would eventually rise up in the coming future, leading to profit gains.Item Machine learning approach of price prediction(AIKTC, 2023-05) Jamkhandikar, Irfan; Mohammad Farhan, (19CO38); Shaikh, Afsar Ahmed (19CO52); Thokan, Naveed Naushad (19CO60)This paper presents a Laptop price prediction system by using the supervised machine learning technique. The research uses multiple linear regression as the machine learning prediction method which offered 85% prediction precision. Using multiple linear regression, there are multiple independent variables but one and only one dependent variable whose actual and predicted values are compared to find precision of results. This paper proposes a system where price is dependent variable which is predicted, and this price is derived from factors like Laptop’s model, RAM, ROM (HDD or SSD), GPU, CPU, IPS Display, and Touch Screen. Price prediction is a useful feature forconsumers as well as businesses. A price prediction tool motivates users to engage with a brand or evaluate offers in order to spend their money wisely. Price prediction enables businesses to set pricing in a manner that builds customer engagement and loyalty. With Machine Learning (ML) technology a price prediction problem is formulated as a regression analysis which is a statistical technique used to estimate the relationship between a dependent/target variable and single or multiple independent (interdependent) variables. In regression, the target variable is numeric. This project will focus on ML algorithm used for price prediction.Item Blockchain based e-voting system(AIKTC, 2023-05) Kazi, Samreen Banu; Ansari, Anas (20DCO02); Bhatkar, Zohaib (20DCO03); Khan, Absar (20DCO04); Shaikh, Anas (20DCO08)Electronic voting or e-voting has fundamental benefits over paper based systems such as increased efficiency and reduced errors. The electronic voting system tends to maximize user participation, by allowing them to vote from anywhere and from any device that has an internet connection. The blockchain is an emerging, decentralized, and distributed technology with strong cryptographic foundations that promises to improve different aspects of many industries. Expanding e-voting into blockchain technology could be the solution to alleviate the present concerns in e-voting. Here we propose a blockchain-based voting system that will limit the voting fraud and make the voting process simple, secure and efficient.