Evaluation Effect Of Color Difference Formulas Based On The YP Color Space

In the production process of color-related industries,it is necessary to strictly control the color quality of products,so the color difference evaluation is particularly important.People have been looking for a color difference formula consistent with the visual characteristics of human eyes.In this paper,378 pairs of color samples are selected from the three directions of hue,chroma and concentration based on the YP color space.First,the RGB values of the selected color samples are obtained by using Datacolor650 spectrometer.Then,the RGB values of color samples and adjacent color samples are mapped on the computer Powerpoint,and the psychophysical methods of gray sacle method is used for visual evaluation.Obtain the visual color difference data of the color sample.Then,the color difference values of selected color samples and adjacent color samples were calculated by CIELAB,CMC(2:1),CIEDE2000(2:1:1),DIN99,CIECAM02-SCD and CIECAM02-UCS,and the relationship between visual color difference and calculated color difference was compared.Finally,using the PF/3(performance factor),the STRESS(Standardized Residual Sum of Squares)and Ftest to evaluate the consistency of the computed chromatic aberration and visual chromatic aberration,and find the color difference formula conforming to the visual characteristics of human eyes.The details are as follows:The reliability of visual data is analyzed through the STRESS(Standardized Residual Sum of Squares)value,that is,the accuracy of the observer is analyzed.The accuracy of the observer includes the uncertainty of the observer itself and the difference between the observers.The results showed that the average STRESS value of 8 observers was 4.7%,the maximum value was 8.8%,and the minimum value was2.3%.The mean STRESS value among 8 observers was 4.3%,with a maximum of 7.6%and a minimum of 3.0%.Both the uncertainty of the observers themselves and the STRESS of the difference between the observers are within the acceptable range,indicating that the visual data can be used for the subsequent research and analysis.At the same time,it was found that the training effect existed in the repeatability results and the accuracy results,that is,with the increase of the number of experiments,the experience of the observers increased,and the experimental data became more real and reliable.Under different hues,all color samples with concentration D=51 and chroma C=57were selected as the central samples,and the average value of visual color differenceΔV of color samples was 1.28.The order of the average color difference ΔE calculated by the six color difference formulas is CIELAB > CIECAM02-UCS > CMC(2:1)>CIECAM02-SCD > CIEDE2000(2:1:1)> DIN99.According to the results of PF/3and STRESS,the CIEDE2000(2:1:1)color difference formula has the smallest PF/3value and the STRESS value,which are 12.9 and 22.7% respectively.Therefore,it can be considered that the CIEDE2000(2:1:1)color difference formula has the best predictive performance.The average six color difference formulas of F-test were used to predict the difference significance of performance.The results show that CIEDE2000(2:1:1)color difference formula is obviously superior to the other five color difference formulas.Therefore,CIEDE2000(2:1:1)color difference formula is selected as the best color difference formula to predict the calculation of color difference of different color samples.Under different chroma,all color samples of different chroma with concentration D=51,yellow(H=0°),orange(H=60°),red(H=120°),purple(H=180°),blue(H=240°)and green(H=300°)were selected.The average value of visual color difference of the six color samples was 1.12.The order of the average color difference ΔE calculated by the six color difference formulas is CIELAB>CIECAM02-UCS>CIECAM02-SCD>CMC(2:1)>DIN99>CIEDE2000(2:1:1).According to the results of PF/3 and STRESS,the prediction performance of different color difference formulas is different.Based on the results of six color systems,CIEDE2000(2:1:1)color difference formulas have the minimum PF/3 value and the STRESS value,which are 13.61 and 26.09%respectively.Therefore,it can be considered that CIEDE2000(2:1:1)color difference formula has the best predictive performance.Under different depth,all color samples with different concentrations of chroma C=24,yellow(H=0°),orange(H=60°),red(H=120°),purple(H=180°),blue(H=240°)and green(H=300°)were selected.The average value of visual color difference of the six color samples was 1.09.The order of the average color difference ΔE calculated by six color difference formulas is CIECAM02-UCS>CIELAB>DIN99>CIECAM02-SCD>CIEDE2000(2:1:1)>CMC(2:1).According to the results of PF/3 and STRESS,the prediction performance of different color difference formulas is different.By combining the results of the six color systems,the PF/3 values of CIECAM02-SCD and CIECAM02-UCS color difference formulas are 7.83 and 7.90,and the STRESS values are 16.02% and 16.21%.Therefore,the prediction performance of the CIECAM02-SCD and CIECAM02-UCS color difference formulas is the best.