Project Title: An Intelligent Object Detection System for Evaluating Assignments in Theory of Automata and Operating System Courses
Category: Web Application
Project File: Download Project File
Muhammad Umar Farooq
umarfarooq@vu.edu.pk
live:umarvc
An Intelligent Object Detection System for Evaluating Assignments in Theory of Automata and Operating System Courses
Project Domain / Category
Web Application.
Abstract / Introduction
Courses such as Theory of Automata and Operating Systems often require assignments containing diagrams such as state machines and system models. Manually grading these assignments is time-consuming and subject to human errors. With the advent of artificial intelligence in academia, the proposed system leverages computer vision and deep learning techniques to accurately analyze and categorize submitted assignments. By automating the detection of objects within diagrams, minimizes human error and significantly accelerates the evaluation process.
An Intelligent Object Detection System for Evaluating Assignments in Theory of Automata and Operating System Courses offers a practical solution by recognizing patterns and evaluating submissions with minimal human intervention.
Functional Requirements:
An Intelligent Object Detection System for Evaluating Assignments in Theory of Automata and Operating System Courses, classify, and evaluate objects such as diagrams and symbols along with written content. It will provide accurate assessment results, constructive feedback, and grading with minimal human intervention.
The proposed system will have the following main users:
Admin Teacher and Student.
Registration module: It will facilitate the registration process for students and teachers. Admin will approve and perform activation of the students and teachers accounts and registration requests.
Login Module: After successful registrations, all types of users will be able to login to the system by using their registered email and password.
Your application will assist the teacher with the scheduling of quizzes and assignments of different subjects including Theory of Automata and Operating System courses.
The system can generate objective and subjective questions of different cognitive levels of Theory of Automata course, including subdomains like: Finite Automata, Generalized Transition graph Regular Languages, and Turing Machines.
The system can generate objective and subjective questions of different cognitive levels of Operating Systems course with subdomains like: Process Management, Memory Management, Deadlock detection through Resource allocation graph (RAG) and File Systems.
Teachers can review and validate a custom-labeled dataset of existing objective and subjective questions, along with their correct answers, for both Theory of Automata and Operating System domains to ensure high evaluation accuracy.
The system aligns assignment grading and evaluation with a predefined rubric, ensuring fairness, accuracy, and consistency in the assessment process.
For Theory of Automata course, the system can leverage advanced computer vision and transformer-based deep learning models such as YOLOv8, Detectron2, OpenCV, GPT-4, LLaMA-
Page 19 of 167
3, Claude, BERT, RoBERTa, DistilBERT, and XLNet to automatically detect, analyze, and evaluate diagram-based questions (e.g., finite automata, state transition diagrams)
In the Theory of Automata course, the system can assign evaluate the structure and intent of diagrams and can convert it to: Finite Automat (FA) → Transition Graph (TG) → Generalize Transition Graph GTG →Regular expression (RE) and vice versa.
The system should evaluate Operating Systems assignments involving diagrams like Resource Allocation Graph (RAG) that is used for deadlock detection.
For the Operating Systems course, the system can use advanced code-evaluation models (e.g., OpenAI Codex, AlphaCode, PolyCoder) to automatically compile/evaluate programming-related tasks such as scheduling, memory management and simulations.
The system should allow teachers to export grades and performance reports in common formats like CSV, Excel, and PDF for record-keeping and analysis.
The system should evaluate the performance of various Deep Learning and Machine Learning models (e.g., OpenCV, YOLO, GPT-4, LLaMA-3, BERT).
For validation purpose, metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R² Score be used to accurately measure model precision and reliability.
The system shall collect each set of student answers and perform automatic evaluation by comparing the model-predicted grades with the actual instructor-assigned (domain expert) grades. Finally applying MSE, RMSE, MAE, and R² Score to accurately assess the model’s grading performance.
The system should allow teachers to define threshold limits for MSE, RMSE, MAE, and R² Score, automatically triggering an alert for manual review if the model’s evaluation metrics exceed the specified limits.
The system should provide teachers with detailed question-level analytics, including the average score, standard deviation, and difficulty index (percentage of students who answered each question correctly) to support in-depth performance evaluation.
The system should automatically identify and flag questions with a high incorrect response rate, suggesting that such questions may be too challenging or ambiguously worded and may need some sort of revision.
The system should allow teachers to generate detailed reports showing student performance on each question and create custom rubrics for evaluating different types of subjective questions that involve diagrams.
The system should provide a help centre with FAQs and live/chat support to assist students with technical issues or questions related to assignments and platform usage.
Tools: JSP, PHP, Python, JavaScript/HTML/CSS, MySQL, PyTorch, TensorFlow, Keras, Pandas, OpenAI Codex, PolyCoder, OpenCV, YOLO, YOLOv8, Faster R-CNN, BERT, GPT-4, LLaMA-3, Claude, RoBERTa, DistilBERT, and XLNet.
Supervisor:
Name: Muhammad Umar Farooq
Email ID: umarfarooq@vu.edu.pk
MS Teams ID: umar.vc@outlook.com
No schedules available for this project.
No reviews available for this project.