Phocus

Inspiration

Have you ever tried to sit down and work on an important assignment, when suddenly DING, you pick up your phone and get stuck in a doom-scrolling vortex for an embarrassing amount of time? Well don't worry, you are not alone. People in everyday life get distracted by their phones, causing serious disruption to their periods of focus and study. We were inspired to try to come up with a solution for this problem.

What it does

Through the use of a webcam, whenever the user picks up their phone, the computer will make an alert for the user to convince them to put their phone away. Additionally, stats for a session are tracked to convince the user to spend less time on their phone and get a new high score.

How we built it

We first trained a object detection model through a custom dataset created in cvat and roboflow. We then trained the model using YOLOv11, and used that model with OpenCV to track the webcam footage.

Challenges we ran into

We struggled with finding the best platforms to use for frontend that easily connects to backend. Simple options had minimal customizability yet the complex ones were too hard to implement given the time. Creating the dataset posed some problems in the correct tracking of the data. This required us to have to retrain the YOLO model until it was working sufficiently.

Accomplishments that we're proud of

We created our own computer vision model, and connected it to a website.

What we learned

A lot of machine learning stuff.

What's next for Phone Detection

Implementing more features focused on helping the user use this alert system to increase their productivity or decrease their phone dependence. Features could include -

• Longer term statistics
  • Average phone time per session
  • Average time spent on phone each pickup
• Setting goals
• Tracking more bad habits
  • Biting/picking nails
  • Picking face/nose
  • Biting their shirt

Built With

• cvat
• opencv
• python
• roboflow
• yolov11