bloom

Building Learning Opportunities Optimized by Models.

Bloom is a e-learning solution that provides teachers with real-time sentiment analysis of their students during online classes.

Inspiration

During the pandemic, online school affected teachers and students dramatically. Lessons were harder to engage with, and students often got distracted or drowsy. Normally, teachers would notice this behavior in the classroom but online learning has hampered the teacher's ability to see everything. Bloom lets teachers get live statistics about how their students are feeling and gives them back their sight through computer vision.

What it Does

With bloom, teachers don't have to worry about focusing on student engagement. Using the power of AI and Computer Vision, teachers get a dashboard of student engagement in real-time. While teaching a lesson, teachers can see the general sentiment in the room and see if they're actually listening or falling asleep. After the class, teacher's can also reflect on their lesson and see which parts of the lesson were more engaging than others.

How we built it

We built this project using a variety of services including Google Compute Engine, Google Cloud Storage, Google AutoML, Firebase, ElectronJS, ChartJS, and YOLOv4. We combined multiple pretrained and trained our own models to return different sentiment scores that be used by teachers including happiness, saddness, tiredness, and engagement.

Accomplishments

As freshmen, we are proud of ourselves for pushing our knowledge and working with a variety of unfamiliar technologies. Understandings how to implement servers, work with Electron, and other AI technologies to return scores were all very difficult and consumed many hours of sleep. We also read through a variety of research papers to learn about different techniques for our task.

Built With

• chartjs
• electron
• firebase
• google-automl
• google-cloud
• google-compute-engine
• yolo