Have you ever wondered if your students are paying attention? understanding the material? feeling bored? Understanding students' emotion in a classroom has always been a challenge for instructors. To improve learning experience for students, it is crucial for instructors to understand parts that students are unclear about.

Our website bridges the communication gap between instructors and students by capturing students' emotions and making a live analysis of the classroom's emotion towards the subject in discussion. By having information on students' reaction to a subject or teaching method, teachers can understand better when they need to modify their teaching methods, or if a certain teaching method is effective.

What it does

Our website can analyze pictures of students in lecture class and detects each student's emotion in the picture. The information is then analyzed and displayed as percentages of students that are happy, bored, or frustrated. We created the algorithm of analyzing students' emotion information and displayed them using visualization libraries. We also created a scoring system based on student's emotion data. Our app assumes that a camera can be installed in classrooms that takes pictures of students while lecture is in session.

How I built it

We used Microsoft azure's emotion api to detect student's emotion in a picture. The emotion information is received as JSON files then processed in python. We wrote an algorithm that detects number of people being happy, neutral and frustrated. Base on this data we give a score of how the teacher is doing. All is then displayed in visualization on the website.

Challenges I ran into

-Microsoft Api takes only pictures through an url, so we ran into various trouble mapping pictures taken through camera to a valid url that azure accepts. -Microsoft api does not detect faces that are not directly facing the camera, so giving generalizations of students' emotions were difficult because we do not know how many people we are detecting out of an entire class.
-We used flask to create the website, and we ran into troubles using flask with visualization libraries.

Accomplishments that I'm proud of

We came up with a lot of ideas to overcome our shortcomings and overcame various technical difficulties.

What I learned

We learned a lot about visualizations and processing image data and creating algorithms to analyze data accurately.

What's next for Classroom Monitor

Integrate this in real classrooms by integrating it as phone applications. Improve accuracy of generalizing students' emotions and scoring algorithm. Implement user profile for teachers so they can track their scores and suggest possible improvements for teaching methods.

Built With

Share this project: