Oclear

project augmented reality ar mixed reality oculus rift

Goal

Apply daltonisation techniques in real-time to an augmented reality view

Key success factors

  • Real-time
  • “Natural looking” - ie. With least amount of distortion to existing colours possible
  • Scores appreciably higher in colour differentiation and colour recognition

The goal is not for the camera images to be accurate, just the transformation.

Methods

Augmented reality

Using video pass-through, an Oculus Rift DK1 with two Playstation Eye cameras mounted.

Daltonisation

Initial tests using the Brettel et al. simulation and daltonisation algorithms were positive, but conclusions led to the simulation not being entirely accurate, thus skewing the final daltonised image unnecessarily. Future algorithms should begin with the simulation techniques detailed by Capilla et al., corresponding-pair procedure, and combining it with the daltonisation techniques of Brettel et al. by using the difference between the original and simulation to distinguish areas of confusion.

Testing

Tests take place using a combination of the Farnsworth-Munsell D15 Color Vision Test, the Ishihara Colour Plate Test and general image comparisons, with and without transformation.

Top: CVD simulation. Bottom: Early isolated daltonisation testing.

Short presentation: Presentation slides at Prezi

Written by Matt Sanders
comments powered by Disqus