ABSTRACT
This project proposes the development of a computer vision system for automatically grading multiple-choice exams. The system will use image processing techniques to extract features from scanned answer sheets and train a machine learning model to classify marks as correct or incorrect. The system will also be able to detect and correct errors in mark making.
The project will be divided into three main phases:
- Data collection: A dataset of scanned answer sheets will be collected from UNIBEN. The dataset will be labeled with the correct answers for each mark.
- Image processing: Image processing techniques will be used to extract features from the scanned answer sheets. These features will be used to train the machine learning model.
- Machine learning: A machine learning model will be trained to classify marks as correct or incorrect. The model will be trained on the dataset of extracted features.
The project will also explore the following areas:
- The use of different computer vision techniques to improve the accuracy of the grading system.
- The development of a user-friendly interface for the grading system.
- The integration of the grading system with other educational systems.
The project is expected to benefit a variety of stakeholders, including students, lecturers, schools, and examination boards. Students will benefit from a faster and more accurate grading process, which will allow them to receive their results more quickly. Lecturers will benefit from a reduction in grading workload, which will allow them to focus on other aspects of their job. Schools will benefit from a more efficient and cost-effective grading process. Examination boards will benefit from a more reliable and consistent grading process.
The project is significant because it has the potential to improve the efficiency, accuracy, and accessibility of exam grading. The project is also innovative because it uses computer vision techniques to automate the gradingprocess.
