This work studies colour constancy in virtual reality under two illuminants using three different methods: selection task, asymmetric matching, and achromatic adjustment. Results indicated that the selection task achieved the highest Color Constancy Index (CCI), highlighting its effectiveness in addressing colour constancy. In contrast, as reported by participants, the asymmetric matching task showed poor CCI performance, largely due to difficulty in achieving precise matches. The achromatic adjustment task showed consistent CCIs across different positions, suggesting stable performance in neutral and coloured illumination conditions.
The indoor scene setup includes a neutral light in the middle of the room and sunlight illumination coming through the glass windows. The 3D environment is rendered using Unreal Engine and visualised using the Varjo Aero headset. Characterization and measurements are conducted to ensure colour accuracy. There is a table in the room positioned to allow objects along it to be illuminated mainly by the outdoor light source, or by the neutral indoor light source.
In summary, the selection task emerged as the most reliable for assessing colour constancy, while other tasks faced usability and precision issues. These findings provide a basis for improving experimental designs to understand colour constancy mechanisms under complex lighting conditions.