Correlations Between Colour Discrimination and Colour Quality Metrics

Nowadays, many studies are trying to propose a measure for characterizing the colour discrimination of human vision. Our previous work conducted three sets of colour discrimination experiments using the Farnsworth -Munsell (FM) 100 Hue Test under lighting conditions with different correlated colour temperatures, illumination levels and Duv values. In this study, the Pearson correlation coefficients between observers’ average error score and 27 typical colour quality metrics were computed based on the three sets of experimental data. The results show that among these metrics, CDI, CSA, GAI, GAI-RA and FSCI have stronger correlations with colour discrimination for most lighting conditions, and relatively, CQI-1, DSI(D65) and Snuetral perform better.