Over the last year, students have been facing many problems with online learning, and some of those problems increased when hybrid learning began. The problem we are trying to solve, is how to bridge the gap between the lack of interaction and easy distraction in online learning, with the interactive, fun environment in a real classroom. To allow for remote learning to be the same as in person, we use a machine learning algorithm that can measure the student’s engagement in the subject. One of the biggest issues with the remote setting is that there are many distractions for students that can take their attention away from the lessons and cause them to fall behind in class. Our idea will measure the class’s engagement and allow the teacher to understand what works and what doesn’t in the online class compared to in-person.
Inspiration Our major inspiration to do this was the problem we faced personally on our learning during this pandemic. Our team consists of participants from India and the US. So irrespective of the country, we shared the common issues over the same. So being students we had to think for ourselves and created this idea.
What it does
Real Class is beneficial for all types of distance education. The platform can be used by any educational institution as a tool for that enables the educators and tutors to record themselves teaching every day. Meanwhile, the student can watch the video anytime between the stipulated period of time, to get their attendance. There can be a minimum percentage of attendance set by the institution to pass or qualify for a subject or course. When a student starts to watch a video, the system uses their camera to track their attention. This is the place where our Machine Learning comes into role. We use the ML model to trace their engagement and attention. This can be achieved by monitoring the student's facial expression, behavior, actions, eyeball movement, and response to the video. To get check their attention and make it more interactive, we add a chatbox, and the tutors can ask questions in between the lessons. This chatbox enables us to get further details about engagement according to response time and response accuracy. Once the student watches the video, a report is sent to Institution or the faculty which contains the attention & engagement analysis and answers made by the student for the questions.
How we built it
The model we built is the very basic one and just a part of our entire idea. A huge thanks to our mentor Mykola Zaitsev for guidance. He introduced us to OpenCV using which we built the model. We also used several other frameworks which are mentioned in the "Built with" column
Challenges we ran into
As we very beginners to programming, we found it hard to the Eco-System. And this is our first hackathon. In the beginning, we found it challenging to understand the Hackathon environment. Later we had a major challenge taking up the idea into a working prototype. But eventually we managed to accomplish the task using tutorials on the Internet to make a similar model that is the backbone of the idea. [P.S The model is just a rough sample of how the idea works. It still needs a major development]
Accomplishments that we're proud of
First of all, we are very happy that we made a program which runs and gives results in real-time. In the beginning of the hackathon none of us knew even basic programming. Plus the fact that our team was formed properly after the start of the hackathon cut us short on time. But in a short time span, we all got along well with the idea and fellow members. We worked as a team in gathering various information was crucial in framing the idea.
What we learned
We learned about some frameworks and libraries of the Python Programming Language and were introduced to Machine Learning concepts.
What's next for Real Class
The next step is to make the model more precise. And to combine all the features we mentioned. And hopefully, we as a team develop a successful prototype.