journal of dentistry 39s (2011) e37–e44 Available online at www.sciencedirect.com journal homepage: www.intl.elsevierhealth.com/journals/jden Dental ceramics: A CIEDE2000 acceptability thresholds for lightness, chroma and hue differences Marı́a del Mar Perez a,*, Razvan Ghinea a, Luis Javier Herrera b, Ana Maria Ionescu a, Héctor Pomares b, Rosa Pulgar c, Rade D. Paravina d a Department of Optics, Faculty of Science, University of Granada, Campus Fuentenueva s/n 18071, Granada, Spain Department of Computer Architecture and Computer Technology, E.T.S.I.I.T. University of Granada, s/n 18071 Granada, Spain c Department of Stomatology, Faculty of Odontology, University of Granada, Campus de Cartuja s/n 18071, Granada, Spain d Houston Center for Biomaterials and Biomimetics & Department of Restorative Dentistry and Biomaterials, The University of Texas Health Science Center at Houston School of Dentistry, 6516 M. D. Anderson Boulevard, Houston, TX 77030, USA b article info abstract Article history: Objectives: To determine the visual 50:50% acceptability thresholds for lightness, chroma Received 22 June 2011 and hue for dental ceramics using CIEDE2000(KL:KC:KH) formula, and to evaluate the formula Received in revised form performance using different parametric factors. 12 September 2011 Methods: A 30-observer panel evaluated three subsets of ceramic samples: lightness subset Accepted 18 September 2011 (jDL0 /DE00j  0.9), chroma subset (jDC0 /DE00j  0.9) and hue subset (jDH0 /DE00j  0.9). A Takagi– Sugeno–Kang Fuzzy Approximation was used as fitting procedure, and the 50:50% acceptability thresholds were calculated. A t-test was used in statistical analysis of the thresholds Keywords: values. The performance of the CIEDE2000(1:1:1) and CIEDE2000(2:1:1) color difference CIEDE2000 formulas against visual results was tested using PF/3 performance factor. Dental ceramics Results: The 50:50% CIEDE2000 acceptability thresholds were DL0 = 2.92 (95% CI 1.22–4.96; Lightness r2 = 0.76), DC0 = 2.52 (95% CI 1.31–4.19; r2 = 0.71) and DH0 = 1.90 (95% CI 1.63–2.15; r2 = 0.88). The Chroma 50:50% acceptability threshold for color difference (DE0 ) for CIEDE2000(1:1:1) was 1.87, whilst Hue corresponding value for CIEDE2000(2:1:1) was 1.78. The PF/3 values were 139.86 for Acceptability thresholds CIEDE2000(1:1:1), and 132.31 for CIEDE2000(2:1:1). Parametric factors Conclusions: There was a statistically significant difference amongst CIEDE2000 50:50% acceptability thresholds for lightness, chroma and hue differences for dental ceramics. The CIEDE2000(2:1:1) formula performed better than CIEDE2000(1:1:1). # 2011 Elsevier Ltd. All rights reserved. 1. Introduction Instruments for color measurement in dentistry, such as spectrophotometers, colorimeters and spectroradiometers, can help overcoming some shortcomings of visual method by bringing accuracy and reducing the chair side time for color matching, communication and reproduction. In addition, color measuring instruments can be a valuable tool in shade verification (quality control). Colorimeters shown good measurement repeatability but they are subject to systematic errors due to edge-loss effect related with sample surface whilst spectrophotometers precisely measures color from reflectance or transmittance data but they are hard to use for in vivo tooth color measurements.1 In this sense, the recent incorporation of spectroradiometers for color measurements in dental research provided accurate and highly repeatable non-contact measurements.2 The development of the CIELAB * Corresponding author at: Office 137, Department of Optics, Faculty of Science, University of Granada, Campus Fuentenueva s/n 18071. Granada, Spain. Tel.: +34 958246164; fax: +34 958248533. E-mail address: mmperez@ugr.es (M.d.M. Perez). 0300-5712/$ – see front matter # 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jdent.2011.09.007 e38 journal of dentistry 39s (2011) e37–e44 color space and the associated DE*ab, have done much to aid in this process and is extensively accepted in dentistry. Recent studies on perceptibility and acceptability color difference thresholds using computer-simulated teeth and dental ceramics suggested that there is a variance in sensitivity with respect to lightness (L*), green–red coordinate (a*) and blue– yellow coordinate (b*) of CIELAB color space.3,4 The dependence of direction of CIELAB color differences on color sensitivity is not influenced by the experimental conditions of measurement. It occurs due to the fact that the formulas which convert CIE1931 colorimetric values into CIELAB L*a*b* coordinates do not adequately capture the perceived color differences.5 Advanced CIELAB-based color difference formulas were introduced to improve the correlation with visual color differences through the implementation of various corrections of the original CIELAB color difference formula.6 Various practical applications of CIELAB, which assumed that CIELAB is a uniform color space, have shown the need of using weighting factors to predict color differences.7 CMC(l:c), CIE94(KL:KC:KH) and CIEDE2000(KL:KC:KH) each employ such weighting factors to adjust the inaccuracies. CIEDE2000 color-difference formula incorporates specific corrections for non uniformity of CIELAB color space (the socalled weighting functions: SL, SC, SH), a rotation term (RT) that accounts for the interaction between chroma and hue differences in the blue region and a modification of the a* coordinate of CIELAB, which mainly affects colors with low chroma (neutral colors) and parameters accounting for the influence of illuminating and vision conditions in color difference evaluation (the so-called parametric factors: KL, KC, KH).8 The parametric factor ratio was proposed as a way to control changes in the magnitude of tolerance judgments and as a way to adjust for scaling of acceptability rather than perceptibility.9 Several authors assumed that texture only affects lightness tolerances but not chroma or hue tolerances,10,11 and therefore the value KL = 2 was proposed.12 It was found that the rotation term (RT), introduced in CIEDE2000 to weight the interaction between chroma and hue differences, is close to zero for the dental color space.13 Recent dental investigations found that CIEDE2000 color difference formula provided better fit than CIELAB formula in the evaluation of color difference, therefore providing better indicators of human perceptibility an acceptability of color differences between tooth colors.2,14 Nonetheless, it seems appropriate to continue studying the CIEDE2000 weighting functions (SL, SC, SH) and parametric factors (KL, KC, KH), which may result in an even better fit with the visual judgments. In addition, no data is available on CIEDE2000 acceptability thresholds for lightness, chroma and hue differences (DL0 , DC0 and DH0 , respectively). These thresholds can lead to a valid and applicable formula to improve the modelling of tooth colored aesthetic materials and, ultimately, patient satisfaction. The specific aims of this study were to determine the visual acceptability thresholds for lightness, chroma and hue for dental ceramics using CIEDE2000(KL:KC:KH) formula and to evaluate the performance of this color difference formula using different parametric factors. The following hypotheses were tested: (i) there were no difference amongst CEIDE2000 50:50% acceptability thresholds for lightness, chroma and hue; (ii) there was no difference in performance of CIEDE2000(2:1:1) and CIEDE2000(1:1:1) in evaluation of color differences of dental ceramics. 2. Materials and methods 2.1. Sample preparation A total of 58 ceramic discs, 14-mm in diameter and 3-mm thick, were fabricated using mixtures of Vita Omega 900, Vitapan 3D-Master opaque powders, and pink, white, and mauve color opaque powders (VITA Zahnfabrik, Bad Säckingen, Germany).14 The surface to be observed for each disc was polished using silica paper (Struers A/S Ballerup, Denmark) sequentially up to 800-grit paper. The range of the color coordinates of the ceramic discs were L0 = 56.09–75.30, C0 = 5.60–28.89 and h0 = 62.16–85.00. All discs were within the color range of central and lateral incisor and canine teeth as it is reflected on a published study.15 The 58 ceramic disks were combined to create a total of 1653 disc pairs, with CIEDE2000 color differences ranging from 0.10 to 10.02 units. Three subsets were defined for visual judgments as follows: lightness subset – pairs of samples that met jDL0 /DE00j  0.9 – where the total color difference is mainly due to the changes in luminance (40 pairs with DL0 ranging from 0.32 to 8.03); chroma subset – pairs of samples that met jDC0 /DE00j  0.9 – where the total color difference is mainly due to changes in chroma (40 pairs with DC0 ranging from 0.99 to 7.89); hue subset – pairs of samples that met jDH0 /DE00j  0.9 – where the total color difference is mainly due to the differences in hue (31 pairs with DH0 ranging from 0.17 to 3.36). As expected, the range of hue variation is significantly smaller since the samples are intended to mimic color of human natural teeth.15 The total number of selected samples was 111. The values of DL0 , DC0 and DH0 for each pair of samples are shown in Fig. 1. 2.2. Color measurement A non-contact SpectraScan PR-704 spectroradiometer (Photo Research, Chatsworth, USA) was used to measure the spectral reflectance of the ceramic disks. This device measure color in a way that matches the geometry of the visual assessments, and has been previously used in dental research.2,16,17 The ceramic discs were placed in the centre of a viewing cabinet (VeriVide CAC60, VeriVide Limited, Leicester, United Kingdom) on a 458 tilted base and a light source simulating the spectral relative irradiance of CIE D65 standard illuminant were employed to provide consistent illuminating/viewing conditions. The discs were positioned 40 cm away from the spectrorradiometer and measured at 08 (corresponding to diffuse/08 illuminating/ measuring geometry). The CIE 1931 28 Standard Colorimetric Observer was used to calculate color coordinate values. Since teeth are translucent and the oral cavity is dark, a Munsell black background (L* = 2.8, a* = 0.7, b* = 1.9) was used for measurements in this study. Similar to previous studies,2,17 a triangular stand was built to hold samples to avoid the specular reflection from the glossy surface. journal of dentistry 39s (2011) e37–e44 DEL ¼ DL0 ðKL  SL Þ DEC ¼ DC0 ðKC  SC Þ DEH ¼ DH0 ðKH  SH Þ e39 where DL0 , DC0 , DH0 are metric differences between the corresponding values of the samples, computed on the basis of uniform color space used in CIEDE2000, and KLSL, KCSC and KHSH are empirical terms used for correcting (weighting) the metric differences to the CIEDE2000 differences for each coordinate. Parametric factors were set to 1 for CIEDE2000(1:1:1) and KL = 2, KH = 1 and KC = 1 for CIEDE2000(2:1:1). When calculating the CIEDE2000 color-difference formula, the discontinuities due to mean hue computation and hue-difference computation as pointed out and characterised by Sharma et al.,19 were taken into account. 2.3. Psychophysical experiment The three subsets (lightness, chroma and hue) of sample pairs were judged by a panel of 30 non-dental professional observers (12 females and 18 males, aged between 19 and 55). All observers were screened for normal color vision using the Ishihara charts (Ishihara Color Vision Test. Kamehara Trading Inc., Tokio, Japan 2004) and all had previous experience in color discrimination experiments. The psychophysical experiment was approved by the Institutional Review Board. During the visual judgments, the observers were positioned approximately 40 cm away from the ceramic disc pairs, which was the same distance used for instrumental color measurements. Each observer was instructed to focus their attention on the centre of the ceramic disks and answer the following question: ‘‘Would you rate the color difference between the two discs as acceptable’’. The responses for each pair of ceramic disks and each observer (DV – visual color difference) were processed.2 2.4. Fig. 1 – DL0 , DC0 and DH0 values of each sample pair of ceramic disks. The CIEDE2000 color difference (DE00) was calculated as follows:6,8 "  2  2  2  #1=2 DL0 DC0 DH0 DC0 DH0 þ þ þ RT DE00 ¼ KL SL KH SH KH SH KC SC KC SC In addition, the CIEDE2000 lightness (DEL), chroma (DEC) and hue (DEH) color differences were defined as follows:18 Fitting procedure A Takagi–Sugeno–Kang (TSK) Fuzzy Approximation model20,21 with Gaussian membership functions and constant consequents was used as fitting method (Matlab 7.1 Fuzzy Logic Toolbox, MathWork Inc., Natick, MA). In the approximations performed, the TSK models took the rule centres equally distributed along the input space, and the rule consequents were optimally obtained using their derivatives with respect to the model output in the minimisation of the value of r (Least Squares LSE approach).22 The number of rules in each case was selected using a 10-fold cross-validation procedure; the number of rules for which the model provided a lowest cross-validation error was chosen to perform the approximation using all data. The 95% confidence intervals (CI, 95% Lower Confidence Limit – LCL and the 95% Upper Confidence Limit – UCL) were estimated and the 50:50% (50% of positive answers and 50% negative answers) thresholds were calculated. The 50:50% point was defined as the difference at which an e40 journal of dentistry 39s (2011) e37–e44 observers would have a 50% probability of making dichotomous judgement, and represents the level of acceptability for these types of judgments. Acceptability thresholds were calculated for the following set of samples: (a) lightness subset (DL0 ), chroma subset (DC0 ) and hue subset (DH0 ); (b) CIEDE2000 lightness subset (DEL), CIEDE2000 chroma subset (DEC) and CIEDE2000 hue subset (DEH); (c) CIEDE2000(1:1:1) and CIEDE2000(2:1:1) for the whole set of 111 disc pairs. 2.5. Statistical analysis A t-test was used to evaluate the differences amongst the lightness, chroma and hue thresholds, both for the metric differences (DL0 , DC0 and DH0 ) and CIEDE2000 lightness (DEL), chroma (DEC) and hue (DEH) color differences, assuming the 50:50% values as normal distributions with variance estimated according to the confidence intervals of the respective fitting curves (SPSS 15.0.1, SPSS, Chicago, USA). To test the performance of CIEDE2000(1:1:1) and CIEDE2000(2:1:1) color difference formulas against visual results, the performance factor PF/3 was calculated.23 This parameter allows statistical comparison of two data sets by means of combining three measures of fit: gamma factor g, CV and VAB.24 The computation of PF/3 is given as: PF 100  ½ðg  1Þ þ VAB þ ðCV=100Þ ¼ 3 3 A PF/3 value of zero indicates perfect agreement between computed and perceived color differences; higher values correspond to worse agreement. From the mathematical point of view, there is no maximum limit for PF/3 values (can be greater than 100%). 3. Results The TSK Fuzzy Approximation fitted curves of the percentage of DV answers (% acceptable) against the instrumentally measured metric differences (DL0 , DC0 , and DH0 respectively) and against the CIEDE2000 lightness, chroma and hue color differences (DEL, DEC, DEH) with their corresponding 95% confidence curves are presented in Figs. 2–4. For DL0 , the determined 50:50% acceptability threshold was DL0 = 2.92 with 95% CI 1.22–4.96 and r2 = 0.76 (Fig. 2a). From the fitted curve of acceptable percentages versus DEL (Fig. 2b), the threshold value found for lightness was 2.86 (95% CI 1.20–4.84; r2 = 0.75). In the case of the DC0 , the 50:50% acceptability threshold was DC0 = 2.52 units, with 1.31–4.19 95% CI and r2 = 0.71 (Fig. 3a). For chroma, from the fitted curve of acceptance percentages against DEC, we calculated a threshold value of 1.34 (95%CI 0.22–2.96, r2 = 0.56) (Fig. 3b). The DH0 value corresponding to 50:50% acceptability threshold was 1.90, with 95% CI 1.63–2.15; r2 = 0.88 (Fig. 4a), whilst for DEH (Fig. 4b), the corresponding threshold was 1.65 (95% CI 1.10–1.87 and r2 = 0.82). Fig. 2 – TSK Fuzzy Approximation fitted curve for visual acceptability in percentages versus (a) DL0 of pairs of ceramic discs; (b) DEL of pairs of ceramic discs. Fig. 3 – TSK Fuzzy Approximation fitted curve for visual acceptability in percentages versus (a) DC0 of pairs of ceramic discs; (b) DEC of pairs of ceramic discs. e41 journal of dentistry 39s (2011) e37–e44 Fig. 4 – TSK Fuzzy Approximation fitted curve for visual acceptability in percentages versus (a) DH0 of pairs of ceramic discs; (b) DEH of pairs of ceramic discs. The t-test confirmed that there were statistically significant differences between threshold values when calculated with the metric differences in lightness (DL0 ), chroma (DC0 ) and hue (DH0 ) ( p < 0.001). For the CIEDE2000 lightness (DEL), chroma (DEC) and hue (DEH), the statistical analysis showed significant differences between DEL and DEC threshold values and between DEL and DEH threshold values, with a negligible pvalue ( p < 0.001). However, when comparing DEC and DEH thresholds, the p-value was higher ( p = 0.1), showing that it is not possible to confirm the difference amongst these two thresholds values. When considering KL = 2, KC = 1 and KH = 1, the values of the DEL, DEC, and DEH thresholds were found to be 1.43, 1.34 and 1.65 respectively, and the statistical analysis showed no significant differences between threshold values of lightness and chroma or between lightness and hue ( p > 0.1 for both). The dependence of the threshold value on the color coordinates is presented in Fig. 5. The percentages of visual acceptability against DL0 and the average value of lightness of each judged pair of samples is presented in Fig. 5a. The visual acceptability percentages of chroma against DC0 and average CIEDE2000 chroma is shown in Fig. 5b, whilst the visual acceptability percentages of hue against DH0 and average CIEDE2000 hue angle for each judged pair of samples is illustrated in Fig. 5c. TSK Fuzzy Approximation fitted curves of the percentage of DV answers (% acceptability) against the instrumentally measured differences CIEDE2000(1:1:1) and CIEDE2000(2:1:1) respectively, with their corresponding 95% confidence curves, are plotted for the three data subsets jointly (Fig. 6). For the CIEDE2000(1:1:1) the determined 50:50% acceptability threshold was 1.87 (r2 = 0.62; Fig. 6a) and optimal number of rules for the TSK Fuzzy Approximation equal to 3. In the case of the CIEDE2000(2:1:1), the 50:50% acceptability threshold was 1.78, with r2 = 0.68 and optimal number of rules for the TSK Fuzzy Approximation equal to 4 (Fig. 6b). Table 1 shows the values of g, CV, VAB and PF/3 for the two color-difference formulas. Lower values of the gamma factor, CV and VAB are obtained in the case of the CIEDE2000(2:1:1) color-diffrence formula, indicating a better adjustment of this formulas to the visual data. 4. Discussion The threshold values should be used in interpretation of clinical and dental laboratory results in terms of acceptability of color differences between natural teeth and dental restoration. Many industries have established industry color tolerances (textile, car industry, money printing) and there is a rising interest to establishing these tolerances for dentistry, especially if using the most advanced methods and means. Practical application of technology that quantifies color and color differences in dentistry requires that color difference formulas provide a quantitative representation of the visual color difference. Recent works2,14,25 reported that it is expected that the CIEDE2000 color difference formula should be used in dental research and other dental applications. It has been reported that the DEab and DE00 might be used interchangeably for the evaluation of color differences in dentistry.26 It should be noted that this interchangeability may be valid for specific region of color space that corresponds to human teeth, but it is not necessarily generally valid. Another study13 reported a significant correlation between DEab and Table 1 – Fit parameters and performance factors for the CIEDE2000(1:1:1) and CIEDE (2:1:1) color-difference formulas against visual judgments for the whole set of samples. Color-difference formula CIEDE2000(1:1:1) CIEDE2000(2:1:1) Factor g CV VAB PF/3 2.98 2.87 88.98 83.03 1.33 1.27 139.86 132,31 e42 journal of dentistry 39s (2011) e37–e44 Fig. 5 – Acceptance percentages against: (a) lightness differences and average lightness of pairs of ceramic discs; (b) chroma differences and average chroma of pairs of ceramic discs; (c) hue differences and average hue angle of pairs of ceramic discs. DE00, even if a significant involvement of weighting functions for lightness, chroma and hue components was observed, thus recommending the use of DE00. A set of color samples of dental ceramic was created for this study. The 50:50% lightness, chroma and hue acceptability thresholds and the CIEDE2000(1:1:1) and CIEDE2000(2:1:1) acceptability thresholds were calculated using TSK Fuzzy Approximation. Data fuzzy modelling represents a flexible and effective method to model an unknown function from a set of observed data relative to that function or phenomenon.20,21 It was previously reported that it enabled soft and accurate approximations, without limiting the expected shape of the objective function, thus allowing the adaptation to unknown shapes on the available data, and being a reliable alternative for the color threshold calculation procedure.2 The first null hypothesis was rejected. Our results have shown differences in sensitivity for changes in lightness (DL0 ), chroma (DC0 ) and hue (DH0 ) in dental color space and the t-test confirmed that these differences were statistically significant. This corresponds to the findings obtained using the CIELAB formula.3,4 Recent experimental results showed that the L* scale gives too large DL* values for lightness differences both for dark and for light samples.27 Therefore, in CIEDE2000, lightness is corrected with a specific weighting function (SL). However, our results (Fig. 5a) suggest no dependence between DL0 and L0 and therefore, the lightness correction of CIEDE2000 does not seem to be significant within the studied area of color space (the dental color space). This result is in agreement with the study that showed that the lightness correction is not statistically significant for color pairs having mainly lightness difference.28 It is well documented in the literature29 that in the Euclidean metric of CIELAB, there is an increase of the tolerance when color differences originated essentially from differences in croma or hue, or when the difference in chroma was very large. Our results suggest no dependence between DH0 and H0 (Fig. 5c), but there was a slight dependence between DC0 and C0 (Fig. 5b). Further studies on this topic, with new experimental data, would be beneficial. CIEDE2000 tolerances for chroma and hue differences are corrected with the specific weighting functions: SC and SH, respectively. The results obtained using TSK Fuzzy Approximation suggested that there were differences in sensitivities between CIEDE2000 lightness, chroma and hue difference: 2.86 in lightness, 1.34 in chroma and 1.65 in hue and the statistical analysis confirmed the differences between lightness and chroma and between lightness and hue. However, there were no statistically significant differences between chroma and hue. This difference between DEL and DEC or DEH could be due to the fact that the parametric factor of lightness (KL) should have a different value than KC or KH. The CIEDE2000 DEL, DEC and DEH were calculated to establish the influence of the weighting functions, thus providing a basis to justify considering the use of KL = 2 instead of KL = 1. The factor PF/3 has been widely used as an indicator for the performance of color-difference formulas in comparison with visual results. The results showed that the factor PF/3 for CIEDE2000(2:1:1) color difference formula had lower values than for the CIEDE2000(1:1:1) color difference formula. Moreover, the 50:50% CIEDE2000(1:1:1) and CIEDE2000(2:1:1) acceptability thresholds calculated using TSK Fuzzy Approximation were similar (Fig. 6), but CIEDE2000(2:1:1) acceptability threshold exhibited slightly better r2-value. These results showed that CIEDE2000(2:1:1) color difference formula performed better than CIEDE2000(1:1:1) in evaluation of acceptability thresholds of dental ceramics. Therefore, the second null hypothesis was rejected. journal of dentistry 39s (2011) e37–e44 e43 Fig. 6 – TSK Fuzzy Approximation fitted curve for visual acceptability in percentages versus DE00(1:1:1) and DE00(2:1:1) of pairs of ceramic discs. The three parametric factors introduced in CIEDE2000 are set to 1 for so-called reference conditions6 which include homogeneous samples. The appearance of a surface (texture) structure may influence color-difference sensation. Some publications have clearly shown the effect of texture on perceived color differences.30,31 In general, it has been assumed that texture affects only lightness tolerance but does not affect chroma or hue tolerance.11,12,32 Our experimental results showed higher values for CIEDE2000 acceptability thresholds for lightness compared to chroma or hue thresholds, which, as mentioned above, might be due to surface texture of the sample. The results of this study suggested considering the use of KL = 2 instead of KL = 1 when evaluating CIEDE2000 lightness difference in dentistry due to the high value of the lightness tolerance compared with chroma and hue tolerances. Considering KL = 2, the threshold value for CIEDE2000 lightness was 1.43 and the statistical test confirmed that there were no significant differences between threshold values of lightness and chroma or hue ( p > 0.1 for both). Nevertheless, further research is needed to test the performance of this color difference formula and its adequacy to accurately predict the perceived color difference in dentistry. In particular, would be of great interest to expand the number of colors and types of materials used (dental resin composites, in vivo teeth, etc.) and comparatively analyse data, thereby evaluating the degree of disagreement between observers and calculated color differences. The parametric factors KL, KC and KH proposed in this study correspond exclusively to color differences in lightness, chroma, or hue, which means that the parametric factors have been computed independently. In the general case, a color difference includes simultaneously lightness, chroma and hue differences; thus the three factors must be combined, and possible interactions should be investigated. 5. Conclusions Within the limitations of this study, it was found that:  Differences in sensitivities between CIEDE2000 50:50% acceptability thresholds for lightness, chroma and hue differences for dental ceramics (2.92, 2.52 and 1.90, respectively), were statistically significant.  The CIEDE2000 formula performed better using KL, KC and KH parametric factors set to 2:1:1 than 1:1:1, which recommends its usage in dental research and in vivo instrumental color analysis. Conflict of interest statement The author declares there are no conflict of interest. Acknowledgments Special thanks to the observers who took part in this experiment. The authors acknowledge funding support from the research projects MAT2009-09795 and SAF2010-20558 of Spanish Ministry of Science and Innovation. references 1. Chu SJ, Trushkowsky RD, Paravina RD. Dental color matching instruments and systems. Review of clinical and research aspects. Journal of Dentistry 2010;38:e2–16. 2. Ghinea R, Pérez MM, Herrera LJ, Rivas MJ, Yebra A, Paravina RD. Color difference thresholds in dental ceramics. Journal of Dentistry 2010;38:e57–64. 3. Lindsey DT, Wee AG. Perceptibility and acceptability of CIELAB color differences in computer-simulated teeth. Journal of Dentistry 2007;35:593–9. 4. Douglas RD, Brewer JD. Acceptability of shade differences in metal ceramic crowns. Journal of Prosthetic Dentistry 1998;79:254–60. 5. Wyszecki G, Stiles WS. Color science: concepts and methods, quantitative data and formulae. New York: John Wiley Press; 2000. 6. CIE Technical Report: Colorimetry. CIE pub. no. 15.3. Vienna, Austria: CIE Central Bureau; 2004. 7. Melgosa M. Testing CIELAB-based color-difference formulas. Color Research and Application 2000;25:49–55. 8. Luo MR, Cui G, Rigg B. The development of the CIE 2000 color difference formula: CIEDE2000. Color Research and Application 2001;26:340–50. e44 journal of dentistry 39s (2011) e37–e44 9. Berns RS. Deriving instrumental tolerances from pass–fail and colorimetric data. Color Research and Application 1996;21:459–72. 10. Steen D, Dupont D. Defining a practical method of ascertaining textile color acceptability. Color Research and Application 2002;27:391–8. 11. Choo S, Kim Y. Effect of color on fashion fabric image. Color Research and Application 2003;28:221–6. 12. Mangine H, Jakes K, Noel C. A preliminary comparison of CIE color differences to textile color acceptability using average observers. Color Research and Application 2005;30:288–94. 13. Perez MM, Saleh A, Yebra A, Pulgar R. Study of the variation between CIELAB (E* and CIEDE2000 color-differences of resin composites. Dental Materials Journal 2007;26:21–8. 14. Wee AG, Lindsey DT, Shroyer KM, Johnston WM. Use of a porcelain color discrimination test to evaluate color difference formulas. Journal of Prosthetic Dentistry 2007;98:101–9. 15. Gozalo-Diaz DJ, Lindsey DT, Johnston WM, Wee AG. Measurement of color for craniofacial structures using a 45/ 0-degree optical configuration. Journal of Prosthetic Dentistry 2007;97:45–53. 16. Perez MM, Saleh A, Pulgar R, Paravina RD. Light polymerization-dependent changes in color and translucency of resin composites. American Journal of Dentistry 2009;22:97–101. 17. Luo W, Westland S, Ellwood R, Pretty I, Cheung V. Development of a whiteness index for dentistry. Journal of Dentistry 2009;37:e21–e26. 18. Nayatani Y. Differences in attributes between color difference and color appearance (chroma an hue) for nearneutral colors. Color Research and Application 2004;29:42–52. 19. Sharma G, Wu W, Dalal EN. The CIEDE2000color-difference formula: implementation notes, supplementary test data, and mathematical observations. Color Research and Application 2005;30:21–30. 20. Takagi T, Sugeno M. Fuzzy identification of systems and its applications to modelling and control. IEEE Transactions on Systems Man and Cybernetics 1985;15:116–32. 21. Herrera LJ, Pulgar R, Santana J, Cardona JC, Guillen A, Rojas I, Perez MM. Prediction of color change after tooth bleaching using fuzzy logic for Vita Classical shades identification. Applied Optics 2010;49:422–9. 22. Herrera LJ, Pomares H, Rojas I, Valenzuela O, Prieto A. TaSe, a Taylor series based fuzzy system Model that combines interpretability and accuracy. Fuzzy Sets and Systems 2005;153:403–27. 23. Guan SS, Luo MR. Investigation of parametric effects using small color differences. Color Research and Application 1999;24:331–43. 24. Huertas R, Melgosa M, Hita E. Influence of random-dot textures on perception of suprathreshold color differences. Journal of the Optical Society of America A 2006;23:2067–76. 25. Johnston WM. Color measurement in dentistry. Journal of Dentistry 2009;37:e2–6. 26. Paravina RD, Kimura M, Powers JM. Evaluation of polymerization-dependent changes in color and translucency of resin composites using two formulae. Odontology 2005;93:46–51. 27. Chou W, Lin H, Luo MR, Westland S, Rigg B, Nobbs J. The performance of lightness difference formulae. Coloration Technology 2001;117:19–29. 28. Melgosa M, Huertas R, Berns RS. Relative significance of the terms in the CIEDE2000 and CIE94 color difference formulas. Journal of the Optical Society of America A 2004;21:2269–75. 29. Melgosa M, Hita E, Perez MM, El Moraghi A. Sensitivity differences in chroma, hue, and lightness from several classical threshold datasets. Color Research and Application 1995;20:220–5. 30. Montag ED, Berns RS. Lightness dependencies and the effect of texture on suprathreshold lightness tolerances. Color Research and Application 2000;25:241–9. 31. Xin JH, Shen HL, Lam CC. Investigation of texture effect on visual color difference evaluation. Color Research and Application 2005;30:341–7. 32. Griffin LD, Sepehri A. Performance of CIE94 for nonreference conditions. Color Research and Application 2002;27:108–15.