Video Tracking for Behaviour Research

Video Tracking for Behaviour Research

Camera and lighting, done right

Nine tracking headaches out of ten start at the camera. Fix the set-up and the software has an easy life.

Where the camera goes

Directly above the arena centre, so the image isn't skewed. Large pools in low rooms may need a wide lens.

Frame rate and resolution

For rodents, 25 to 30 frames a second is usually plenty; fast behaviours and larval fish may need more. Aim for the animal to be at least a few dozen pixels across.

Light

  • Even, diffuse light with no hotspots on the water.
  • Infrared lights and an IR-sensitive camera for dark-phase tests.
  • Identical lighting across animals and days. It changes behaviour as well as tracking.

Contrast

Match the arena to the animal: white floors for dark animals, dark floors for albino ones. In the pool, opaque water hides the platform and helps the tracker in one go.

A 60-second check before every session

  1. Is the whole arena in view?
  2. Is the calibration still right?
  3. Does a quick test track run with no lost frames?

Further reading