S ScienceFairProjects.info
Medium⏱ time1.5 hr$ cost0

Webcam Object Detector

Run a pre-trained model that names and boxes objects in your camera view live.

Start building ↓
fig — working model Result first

The build

Step 01

Load the model

Bring in a pre-trained object detector.

Step 02

Stream video

Feed webcam frames to the model.

Step 03

Draw labels

Box and name each detected object.

Step 04

Explore

Point it at things around you.

Working Principle

The pre-trained model recognises dozens of common objects and returns each one's label and location in the frame.

The science behind it

A closer look

Object detection finds and classifies many items at once; the model was trained on a huge labelled image set.

Take it further

Variables to test

  • 1 Which objects does it miss most?
  • 2 Hold two objects — does it find both?

More AI