Inspiration

Our team created a stress analyzer to evaluate people's stress levels. Due to the pandemic, people's life has been changed. Lots of people begin to work from home and are unable to connect with others. Stress comes from work, relationships and everywhere. We want to use our model to help people better understand their stresses in this special time.

What it does

The stress analyzer is embedded in a website we created. The Stress analyzer captures your video using a webcam to analyze your face and posture and displays whether it is stressed or unstressed.

How we built it

  • Collect image data and label samples : we captured around 1000 pictures of ourselves in various positions in front of our webcams as we were not able to find a good dataset
  • Trained the model : We used Google's Teachable Machine to train the data model
  • Developed a website : we used HTML/CSS and JavaScript to integrate the model into our website

Challenges we ran into

One of the challenges was trying to find a good data set from the web before we decided to just use our own photos. Training the model with accurate data was a major challenge and the time taking one.

Accomplishments that we're proud of

We are proud of our team work, effort we put into the project, everything that we learned and how our project turned out.

What we learned

  • 2 out 4 of us had never worked with machine learning before and had never trained a model. We learned how to effectively collect a suitable data set, and also integrate the trained model into our project.
  • Planning and Time management: As we are from different time zones, we discussed and planned.

What's next for Stress Analyzer

  • Upgrading the binary classification model to multiclass classification to analyze different stress level
  • Creating a mobile app for Stress Analyzer
  • A dashboard for user
  • Time series based report and assessment for users
  • Deploying the website so everyone can access it

Built With

Share this project:

Updates