InkGrader

Inspiration

Teaching assistants, professors, and educators often find themselves mired in the enormous task of grading multiple-choice or short-answer quizzes for large classes, diverting their focus from creating an engaging learning environment. The tedious process not only demands a significant amount of time and effort but also bears the risk of human error. Our team was inspired by the struggles faced by educators and the potential of leveraging technology to alleviate this challenge, thereby crafting an efficient, automated, and scalable solution - the "Ink Grader".

What it does

Ink Grader automates the process of grading student quizzes by leveraging image processing, machine learning, and artificial intelligence. Educators simply need to upload images of student answer sheets. The system then employs Optical Character Recognition (OCR) to extract text from the images, utilizes the OpenAI API to comprehend the answers, and compares them against the provided answer key. Consequently, it calculates grades accurately and swiftly, facilitating rapid and consistent feedback to students while substantially reducing the grading workload for educators.

How we built it

The project is developed using Django for the backend. The OCR functionality was implemented using Google Cloud Vision API, which extracts text from the images of the answer sheets. For understanding and analyzing short answers, we integrated the OpenAI API, utilizing its advanced natural language processing capabilities. We also established a database to store, retrieve, and manage the quizzes, answer keys, and respective grades. The front end was crafted using a user-friendly design to ensure ease of use for educators.

Challenges we ran into

Accuracy of OCR: Achieving high accuracy in text extraction from diverse answer sheets and handling various handwriting styles was challenging.

Answer Analysis: Designing a system that can accurately compare student responses with answer keys, especially for non-multiple-choice answers, required innovative solutions.

User Interface: Creating a user-friendly and intuitive interface that catered to the varied tech proficiency of educators demanded careful and considerate design.

Accomplishments that we're proud of

Automation of Grading: Successfully automating a traditionally manual and labor-intensive task, making a tangible positive impact on educators' workload.

Accuracy: Achieving a high level of accuracy in grading, reducing the scope of grading errors and biases.

What we learned

Technical Proficiency: Enhanced our expertise in Django, cloud computing, and integrating various APIs cohesively, despite our teams not having deep levels of experience in these areas.

Project Management: Enhanced our skills in managing diverse aspects of project development, from conceptualization to implementation.

What's next for InkGrader

User Population: Working with professors from our university to help them streamline their grading processes, and collaborate with TAs to see if it helps their lives.

Adaptive Learning: Leveraging the data to provide insights into student performance and enable adaptive learning experiences.

Mobile App: Developing a mobile application to allow educators to use Ink Grader on the go, making the grading process even more convenient.

Data Analytics: Incorporating data analytics for educators to gain insights into class and individual student performance, helping to tailor teaching approaches accordingly.