Powerpoint is ON GITHUB!!!

Inspiration

The likelihood of depression in such a high-stress working environment like ours is likely to surge. The abundance of audio and facial data from remote working, especially more so in the future borne out of the Smart Nation initiative, provides us the opportunity to identify at-risk individuals with the power of data.

What it does

Monitors and identifies personnel likely to be facing burnout and depression through computer vision and deep learning techniques.

How we built it

Captilizes on our combined knowledge in various fields in machine learning and deep learning, from simple models like logistic regression to the more sophisticated ones such as deep generative models (e.g. GANs) and XGBoost.t

Challenges we ran into

Integration of different components into a seamless pipeline.

Accomplishments that we're proud of

The novelty of leveraging big data for preventative healthcare.

What we learned

A lot! From technical skills, such as the techniques involved in the models themselves and learning to build Telegram bots, and organizational skills, like managing a Github repository better in a big group.

What's next for Depressed Detective

Incorporation of more variables such as MEL spectrum frequencies (audio data), and gait analysis. Refine our data collection process and roll out into a testing phase.

Built With

• javascript
• python