## Abstract

## Objectives

To analyze the influence of lightness, chroma and hue differences on visual shade matching performed by dental students.

## Methods

100 dental students (DS) volunteers with normal vision participated in the study. A spectroradiometer (SP) was used to measure the spectral reflectance of 4 extracted human upper central incisors (UCI 1-4) and shade tabs from Vita Classical (VC) and Vita Toothguide 3D-Master (3D) shade guides. Measurements were performed over a gray background, inside a viewing booth and under D65 illuminant (diffuse/0° geometry). Color parameters ( *L **, *a **, *b **, *C ** and *h *°) were calculated. DS used VC and 3D to visually select the best shade match for each UCI. CIE metric differences ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='Δa*,Δb*,ΔL′’>Δa∗,Δb∗,ΔL′Δa*,Δb*,ΔL′

Δ a * , Δ b * , Δ L ′

, <SPAN role=presentation tabIndex=0 id=MathJax-Element-2-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-3-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

) and CIEDE2000(2:1:1) lightness ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-4-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEL’>ΔELΔEL

Δ E L

), chroma ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-5-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEC’>ΔECΔEC

Δ E C

) and hue ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-6-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEH’>ΔEHΔEH

Δ E H

) differences were obtained from each UCI and the first three shades selected by DS and the first option using CIELAB, CIEDE2000(1:1:1) and CIEDE2000(2:1:1) color difference metrics. The closest CIELAB color-discrimination ellipsoid (from RIT-DuPont visual color-difference data) to each UCI was selected for the analysis of visual shade matching.

## Results

DS showed a preference for shades with lower chroma ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-7-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-8-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEC’>ΔECΔEC

Δ E C

) and/or hue ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-9-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-10-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEH’>ΔEHΔEH

Δ E H

) values instead of shades with lower lightness values ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-11-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL′’>ΔL′ΔL′

Δ L ′

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-12-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEL’>ΔELΔEL

Δ E L

). Most best visual matches were near the tolerance ellipsoid centered on tooth shade.

## Significance

This study is an attempt to partially explain the inconsistencies between visual and instrumental shade matching and the limitations of shade guides. Visual shade matching was driven by color differences with lower chroma and hue values.

## 1

## Introduction

Most studies of color difference in dentistry use CIELAB color space and the associated <SPAN role=presentation tabIndex=0 id=MathJax-Element-13-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEab*’>ΔE∗abΔEab*

Δ E a b *

as the standard parameter for total color difference between a pair of colored samples. The greater the value, the larger the color difference and, consequently, the more perceptible the difference to the human eye . CIELAB color differences <SPAN role=presentation tabIndex=0 id=MathJax-Element-14-Frame class=MathJax style="POSITION: relative" data-mathml='(ΔEab*)’>(ΔE∗ab)(ΔEab*)

( Δ E a b * )

are easily computed as the Euclidean distance to measure total distance within this color space, or as the result of the differences in each of the three color attributes ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-15-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL*’>ΔL∗ΔL*

Δ L *

, <SPAN role=presentation tabIndex=0 id=MathJax-Element-16-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC*’>ΔC∗ΔC*

Δ C *

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-17-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH*’>ΔH∗ΔH*

Δ H *

, respectively, lightness, chroma and hue differences) . It has been indicated since 1976 that the CIELAB color space is not uniform and that “in different practical applications it may be necessary to use different weightings for <SPAN role=presentation tabIndex=0 id=MathJax-Element-18-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL*’>ΔL∗ΔL*

Δ L *

, <SPAN role=presentation tabIndex=0 id=MathJax-Element-19-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC*’>ΔC∗ΔC*

Δ C *

, and <SPAN role=presentation tabIndex=0 id=MathJax-Element-20-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH*’>ΔH∗ΔH*

Δ H *

” . Thus, four advanced color difference formulas using weighting factors were developed to predict color differences: CMC <SPAN role=presentation tabIndex=0 id=MathJax-Element-21-Frame class=MathJax style="POSITION: relative" data-mathml='(l:c)’>(l:c)(l:c)

( l : c )

, BFD <SPAN role=presentation tabIndex=0 id=MathJax-Element-22-Frame class=MathJax style="POSITION: relative" data-mathml='(l:c)’>(l:c)(l:c)

( l : c )

, CIE94 <SPAN role=presentation tabIndex=0 id=MathJax-Element-23-Frame class=MathJax style="POSITION: relative" data-mathml='(KL:KC:KH)’>(KL:KC:KH)(KL:KC:KH)

( K L : K C : K H )

and CIEDE2000 <SPAN role=presentation tabIndex=0 id=MathJax-Element-24-Frame class=MathJax style="POSITION: relative" data-mathml='(KL:KC:KH)’>(KL:KC:KH)(KL:KC:KH)

( K L : K C : K H )

. Nowadays, CIE recommends the use of CIEDE2000 color difference formula, even though in the past the CIE94 and CMC formulas were widely used. The parametric factors <SPAN role=presentation tabIndex=0 id=MathJax-Element-25-Frame class=MathJax style="POSITION: relative" data-mathml='KL’>KLKL

K L

= 1, <SPAN role=presentation tabIndex=0 id=MathJax-Element-26-Frame class=MathJax style="POSITION: relative" data-mathml='KC’>KCKC

K C

= 1 and <SPAN role=presentation tabIndex=0 id=MathJax-Element-27-Frame class=MathJax style="POSITION: relative" data-mathml='KH’>KHKH

K H

= 1 (corresponding to lightness, chroma and hue, respectively) were used for computation with the CIEDE2000(1:1:1) color difference formula . However, studies on visual judgments performed on textile color acceptability , dental ceramics acceptability and the comparison of visual and instrumental shade matching in dentistry showed that the use of a parametric factor for lightness <SPAN role=presentation tabIndex=0 id=MathJax-Element-28-Frame class=MathJax style="POSITION: relative" data-mathml='KL’>KLKL

K L

= 2, using CIEDE2000(2:1:1), resulted in color differences that better correlated to visual observations from average observers.

A previous study on dental color space showed differences in sensitivity for changes in lightness ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-29-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL′’>ΔL′ΔL′

Δ L ′

), chroma ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-30-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

) and hue ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-31-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

). These threshold values should be used in the interpretation of clinical results in terms of acceptability of color differences between natural teeth and ceramic restorations. In addition, some formulas such as BFD and CIEDE2000, contemplate the rotation of experimental chromatic discrimination ellipses in the plane *a ** *b ** (color discrimination ellipses predicted by CIE94 and CMC are radially oriented in this plane) .

The main purpose of a color difference formula is to achieve the greatest agreement between the resulting shade from the smallest color difference and the visual judgments from average observers. The just-noticeable color differences or thresholds are usually represented by ellipses/ellipsoids in a specific color space . The nineteen CIELAB color-difference centers (named ellipsoids), originating from RIT-DuPont visual color-difference data, would be helpful to visualize systematic patterns of variation in color-difference behavior, allowing to identify no perceptible points within a specific area. These ellipsoids were useful for the development of color-difference equations based on CIELAB color space .

The purpose of this study was to evaluate the influence of lightness, chroma and hue differences on visual shade matching performed by dental students (DS), using both, the metric differences ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-32-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL′’>ΔL′ΔL′

Δ L ′

, <SPAN role=presentation tabIndex=0 id=MathJax-Element-33-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-34-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

) and CIEDE2000 lightness ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-35-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEL’>ΔELΔEL

Δ E L

), chroma ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-36-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEC’>ΔECΔEC

Δ E C

) and hue ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-37-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEH’>ΔEHΔEH

Δ E H

) differences. In addition, the CIELAB ellipsoids, from RIT-DuPont visual color-difference data, were used to evaluate the visual color differences of chosen shades. Therefore, this study tested the following hypotheses: (1) the independent evaluation of color difference coordinates influence on visual shade matching performed by DS, and (2) the visually perceived shade for each UCI is greater than the color difference thresholds (CIELAB color discrimination ellipsoids).

## 2

## Material and methods

This study was approved by the local Ethics in Research Committee and is registered in the national platform for ethically approved studies involving human beings (# 767.915).

## 2.1

## Instrumental and visual shade matching assessment setup

A spectroradiometer (SP- SpectraScan® PR-704, Color Research Inc., Chatsworth CA, USA) was used to measure the spectral reflectance of 4 extracted human upper central incisors (UCI), 16 shade tabs from Vita Classical shade guide (VC) and 26 shade tabs from Vita Toothguide 3D-Master shade guide (3D) (VITA Zahnfabrik, Bad Säckingen, Germany). These measurements were performed inside a viewing booth (CAC 60, VeriVide Limited, Leicester, United Kingdom) with a gray background (Flexipalette Color Match) and under CIE D65 illuminant (diffuse/0° illuminating/measuring geometry). Color parameters (CIE *L **, *a **, *b **, *C ** and *h *°) were calculated according to the CIE D65 Standard illuminant and the CIE 2° Standard Observer .

Short-term repeated (3×) measurements without replacement were performed in each specimen. Color measurements were performed in the middle third of all UCIs and shade tabs. As for previous studies , a stand was built to support the specimens and avoid specular reflection from the glossy surface. Since the spectroradiometer field of measurement is 1°, it was placed 35 cm from the specimen, allowing for the measurement of the whole specimen.

Visual shade matching performed by a panel of 100 dental students (DS), 35 males and 65 females with normal vision (tested using Ishihara’s Test for Colour-Blindness), with previous knowledge and clinical experience in shade matching were selected to participate in the study. Nevertheless, brief instructions on the use of shade guides were given. Two shade guides: Vita Classical (VC) and Vita Toothguide 3D-Master (3D) (VITA Zahnfabrik, Bad Säckingen, Germany) were used for shade matching. Experimental procedures were performed as previously described .

## 2.2

## Color coordinates differences

Differences in lightness ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-38-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL′’>ΔL′ΔL′

Δ L ′

), chroma ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-39-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

) and hue ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-40-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

) were obtained from CIEDE2000 color difference metric. The CIEDE2000(K _{L }:K _{C }:K _{H }) color differences for lightness <SPAN role=presentation tabIndex=0 id=MathJax-Element-41-Frame class=MathJax style="POSITION: relative" data-mathml='(ΔEL)’>(ΔEL)(ΔEL)

( Δ E L )

, chroma <SPAN role=presentation tabIndex=0 id=MathJax-Element-42-Frame class=MathJax style="POSITION: relative" data-mathml='(ΔEC)’>(ΔEC)(ΔEC)

( Δ E C )

and hue <SPAN role=presentation tabIndex=0 id=MathJax-Element-43-Frame class=MathJax style="POSITION: relative" data-mathml='(ΔEH)’>(ΔEH)(ΔEH)

( Δ E H )

were defined as follows :

Δ E L = Δ L ′ ( K L . S L )

Δ E C = Δ C ′ ( K C . S C )

Δ E H = Δ H ′ ( K H . S H )

where <SPAN role=presentation tabIndex=0 id=MathJax-Element-47-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL′’>ΔL′ΔL′

Δ L ′

, <SPAN role=presentation tabIndex=0 id=MathJax-Element-48-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-49-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

are metric differences between the corresponding coordinate values of each tooth and each one of the shade tabs from VC and 3D shade guides, computed on the basis of uniform color space used in CIEDE2000. The empirical terms <SPAN role=presentation tabIndex=0 id=MathJax-Element-50-Frame class=MathJax style="POSITION: relative" data-mathml='KL.SL’>KL.SLKL.SL

K L . S L

, <SPAN role=presentation tabIndex=0 id=MathJax-Element-51-Frame class=MathJax style="POSITION: relative" data-mathml='KC.SC’>KC.SCKC.SC

K C . S C

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-52-Frame class=MathJax style="POSITION: relative" data-mathml='KH.SH’>KH.SHKH.SH

K H . S H

are used for correcting (weighting) the metric differences to the CIEDE2000 differences for each coordinate.

Parametric factors were set to <SPAN role=presentation tabIndex=0 id=MathJax-Element-53-Frame class=MathJax style="POSITION: relative" data-mathml='KL’>KLKL

K L

= 2, <SPAN role=presentation tabIndex=0 id=MathJax-Element-54-Frame class=MathJax style="POSITION: relative" data-mathml='KH’>KHKH

K H

= 1 and <SPAN role=presentation tabIndex=0 id=MathJax-Element-55-Frame class=MathJax style="POSITION: relative" data-mathml='KC’>KCKC

K C

= 1 for CIEDE2000(2:1:1) . When calculating the CIEDE2000 color-difference formula, the discontinuities due to mean hue computation and hue-difference computation were taken into account, whereby both were pointed out and characterized by Sharma et al. .

Color coordinate differences were finally evaluated through comparisons with 50:50% acceptability thresholds (AT) for lightness, chroma and hue in dental color space. The AT values considered in this study were <SPAN role=presentation tabIndex=0 id=MathJax-Element-56-Frame class=MathJax style="POSITION: relative" data-mathml='ΔL′’>ΔL′ΔL′

Δ L ′

= 2.92, <SPAN role=presentation tabIndex=0 id=MathJax-Element-57-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

= 2.52 and <SPAN role=presentation tabIndex=0 id=MathJax-Element-58-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

= 1.90, for the metric differences and <SPAN role=presentation tabIndex=0 id=MathJax-Element-59-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEL’>ΔELΔEL

Δ E L

= 1.43, <SPAN role=presentation tabIndex=0 id=MathJax-Element-60-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEC’>ΔECΔEC

Δ E C

= 1.34 and <SPAN role=presentation tabIndex=0 id=MathJax-Element-61-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEH’>ΔEHΔEH

Δ E H

= 1.65 for CIEDE2000 lightness, chroma and hue differences, respectively, as previously determined using TSK Fuzzy Approximation .

## 2.3

## The CIELAB color-discrimination ellipsoids

The nineteen color-discrimination ellipsoids in CIELAB color space, originating from RIT-DuPont visual color-difference data, were obtained from a previous study . Only four CIELAB color-discrimination ellipsoids (MG—medium gray, LB—light brown, MY—moderate yellow and LG—light gray) were based on the approximation to the 4 UCIs. The minimum distance (Euclidean distance) between each UCI and the center of these four ellipsoids determined which ellipsoid was used for shade analysis of each tooth. The first three shades chosen from visual matching performed by DS and the first option of each one of the three color difference formulas (CIELAB, CIEDE2000(1:1:1) and CIEDE(2:1:1)) obtained by SP measurements were used to the shade analysis of each UCI. In order to understand the importance of the ellipsoids for visual perception, the translation of the chosen ellipsoid, using CIELAB color coordinates from each UCI as its center point, was performed.

## 2.4

## Statistical analysis

The normality and homogeneity of variance were satisfied by Levene’s test ( *α *= 0.05). Thus, the CIE metric differences ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-62-Frame class=MathJax style="POSITION: relative" data-mathml='Δa*,Δb*,ΔL′’>Δa∗,Δb∗,ΔL′Δa*,Δb*,ΔL′

Δ a * , Δ b * , Δ L ′

, <SPAN role=presentation tabIndex=0 id=MathJax-Element-63-Frame class=MathJax style="POSITION: relative" data-mathml='ΔC′’>ΔC′ΔC′

Δ C ′

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-64-Frame class=MathJax style="POSITION: relative" data-mathml='ΔH′’>ΔH′ΔH′

Δ H ′

) and CIEDE2000(2:1:1) differences ( <SPAN role=presentation tabIndex=0 id=MathJax-Element-65-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEL,ΔEC’>ΔEL,ΔECΔEL,ΔEC

Δ E L , Δ E C

and <SPAN role=presentation tabIndex=0 id=MathJax-Element-66-Frame class=MathJax style="POSITION: relative" data-mathml='ΔEH’>ΔEHΔEH

Δ E H

) between each UCI and all possible shade matches obtained with DS (the first three options) and SP (the first option using each one of the three color difference metrics) using VC and 3D shade guides, were statistically analyzed using one-way analysis of variance (ANOVA) and Tukey’s multiple comparison test with Bonferroni correction, considering the shade tabs and UCI as factors for each parameter ( *p *≤ 0.01). All statistical analyses were performed using a standard statistical software package (SPSS 16.0, Chicago, IL, USA).