journal of dentistry 40s (2012) e3–e10

Available online at www.sciencedirect.com

journal homepage: www.intl.elsevierhealth.com/journals/jden

Dental shade matching using a digital camera

W.K. Tam, H.J. Lee *

Department of Medical Informatics, Tzu Chi University, No. 701, Sec. 3, Jhongyang Rd., Hualien City, Hualien County 97004, Taiwan, ROC

article info abstract

Article history: Objectives: Digital cameras could be substitutes for contact-type instruments in shade
Received 23 June 2011 selection and overcome their drawbacks. The images taken show morphology and color
Received in revised form texture of teeth. A new method was proposed to compare the color of shade tabs taken by a
27 May 2012 digital camera using appropriate color features.
Accepted 9 June 2012 Methods: Vita 3D-MASTER shade guide and Canon EOS 1100D digital camera were employed.
Shade tab images were compared in two reference strategies. The color of tooth surface was
presented by a content manually cropped out of the image. The content was divided into
Keywords: 10  2 blocks to encode the color distribution. Color features from commonly used color
Dental shade matching spaces were evaluated. The top n matches were selected when the least n shade distances
Shade content between the shade tabs were attained.
Tooth color block Results: Using Sa*b* features, the top one accuracy was 0.87, where the feature S is defined in
Shade distance HSV color space, a* and b* features are defined in L*a*b* color space. This rate was higher than
Digital shade matching previous reports using contact-type instruments. The top three matching accuracy was 0.94.
Camera in dentistry Conclusions: Sa*b* were suitable features for shade matching using a digital cameras in this
study. Both the color and texture of the tooth surface could be presented by the proposed
content-based descriptor. Clinical use of digital cameras in shade matching became possible.
Clinical significance: This in vitro study proposed a method for shade matching using digital
cameras through the comparisons of the color patterns on the shade tab surfaces. The
method overcame some drawbacks from the devices such as colorimeters or spectro-
photometers. The results supported the use of digital cameras in shade matching.
# 2012 Elsevier Ltd. All rights reserved.

verbal communication of color differences is limited.3 A good

1. Introduction
Shade selection in dental practice is an important but difficult
task.1,2 Dentists were challenged to satisfy the aesthetic
requirement from patients when they select the shades to
match the natural teeth during fabrication of prosthesis. Most
dentists are usually not well-trained for shade selection in
their daily matching procedures because they lack knowledge
in color science. Though there are many kinds of shade guides
available in the dental market, it is still difficult to match shade
tabs with teeth in the intra-oral environment.
In the procedure of prosthesis fabrication, dentists and
technicians need to communicate on teeth colors. However,
shade matching result as well as color communication directly
relates to the quality of prosthesis. The more precise dental
colors can be described, the more accurate porcelain colors
can be delivered.
Dental colors can be analysed by two methods: visual and
instrumental.4 Traditionally, dentists usually select suitable
shade tabs by their naked eyes.5 Though detailed procedures
of matching have been emphasised on different clinical
settings,6–8 the results were still unreliable and inconsistent.
Most currently available dental products employ color
science to devise a standard for digital descriptions of colors.9–
15
These products, such as computerised colorimeters and
spectrophotometers, have been commercialised for couple of

* Corresponding author. Tel.: +886 3 8565301x2400/2402; fax: +886 3 8579409.

E-mail addresses: tam@ivydental.com.tw (W.K. Tam), hjlee0301@hotmail.com, hjlee@mail.tcu.edu.tw (H.J. Lee).
0300-5712/$ – see front matter # 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jdent.2012.06.004
e4 journal of dentistry 40s (2012) e3–e10

years and proven to have stable results but do not associate uses RGB color space since 1940.31–34 In 1931, the Commission
with higher accuracy.15–17 Most of these instruments employ Internationale de l’E’clairage establishes the XYZ color space
DE of L*a*b* color space, according to the regulation announced composed of tristimulus values representing the human
by American Dental Association (ADA),18 for measuring color vision response to a given color.35 Though all these color
differences of dental shades, which is defined as19: spaces are thought to be consistent, some of their features
1=2
DE ¼ ððLi  Lj Þ2 þ ðai  aj Þ2 þ ðbi  bj Þ2 Þ ; may be less influenced by illuminant variations and can be
used independently in digital shade matching.
where i and j denote two dental shades. The aims of this in vitro study include (1) to overcome the
Clinical observations have claimed that a DE value greater drawbacks of computerised shade matching instruments in
than 3.7 is graded as a poor match in dental shades.15 The uses measuring tooth surface, (2) to define color descriptors of teeth
of these contact-type instruments may fall into a shortcoming based on content-based features, and (3) to select and
that the small (usually 3 mm) window of the devices may not reassemble effective color features under unstable illuminant
capture reliably the real color information of the whole tooth conditions.
surface. It poses difficulties in correctly positioning the device
over the tooth surface during the measuring procedure. The
second shortcoming is the improper measurement of the 2. Methods
curved translucent surfaces found on teeth using the contact-
type instruments that are designed for flat surfaces.20 The 2.1. System flow
third shortcoming is that most of these equipments employ
mean values or intensity histograms for representing the A system was proposed for an in vitro shade matching using a
shade zone for color labelling and ignore the geometric digital camera. The system consisted of three major parts:
distribution of color within the region. Undoubtedly, shade image preprocessing, color features extraction and similarity
determination by visual observation is inconsistent, but the measurement, as depicted in Fig. 1. In the first part, a shade
accuracy of colorimeters in matching shades is only slightly content cropped from the image was segmented into small
better, the matching rate of 50% by the colorimeters and 48% blocks and noise was removed. Second, color features were
by visual selection.16 extracted and reassembled into useful feature sets. Third, test
Recently, the usage of mega-pixel digital cameras have regions were matched with the shade samples in the database
been widely adopted in the fields of dentistry.21 During using the content-based color measurements.
communication with technicians, dentists can present not
only the dental morphology and colors, but also the surface 2.2. Images collection
texture, color distribution and other information under the
intra-oral conditions.22 Intra-oral images with the reference A commercial dental shade guide (3D-MASTER, Vita Zahnfab-
shade tabs correctly positioned next to the teeth are also rik, Bad Sackingen, Germany) was employed in this study,
useful in shade matching.6,23 Shade matching using the digital which has been reported as a reliable guide for instrumental
images can minimise the gap of color communication analysis.36 Shade tabs were labelled in numeric order, from 1
between dentists and technicians. However, unstable illumi- to 26, as shown in Fig. 2.
nation in the images poses a big problem in intensity The images of the shade guide were captured by a digital
normalisation during shade analysis. Meanwhile, the accura- camera (Canon EOS 1100D, Canon, Taiwan) equipped with a
cy of tooth color matching other than measuring DE using ring flash (Sunpak auto 16R pro, SEA&SEA SUNPAK, Japan)
camera images mimicking clinical environment has not been
[(Fig._1)TD$IG]
reported before.
Another problem is concerned with the color description of
tooth surface from an image. A color description should be
able to describe in detail the color distribution of the tooth
surface. It is reported that color alternates along the
longitudinal axis of the tooth surface.24 The gradation starts
from the cervical, which is the most saturated, to the incisal
area.25,24 Measurement of incremental contents of teeth has
been proposed by Shigemi et al. for computerised color
reproducibility of prosthesis.26 The areas along the proximal
contacts with neighbouring teeth and near gingiva are
subjective to change with shadow effect and cannot reveal
a natural tooth color. The central area chosen for matching
records the complexity of color gradation on natural teeth and
can be used as an effective content for shade comparison.
The features in L*a*b* color space, which is defined in 1976
based on the color receptors of human eyes,27 and HSV color
spaces28–30 are commonly used in dentistry. Tooth color
matching may also take advantage of other color spaces
proposed by scientists. The most popular VGA monitor system Fig. 1 – System-flow diagram.
 journal of dentistry 40s (2012) e3–e10 e5
[(Fig._2)TD$IG]

Fig. 2 – Commercial Vita 3D-MASTER dental shade guide with numeric labels.

which emitted light at color temperature of 5500–5600 K. The
capturing environment was defined as that the shade guide
position and the camera shooting distance were usually
adopted by a dentist in a dental clinic. Since dentists seldom
fix the distance between the teeth and camera before taking
an image, there exists variation of shooting distances in
individual shoot. The unstable environmental illumination
and unfixed camera distance were considered as the uncon-
trolled environmental factors in the experiments. In this way,
the shooting distance was set in a range of 30–50 cm that
allowed the whole shade guide being well captured inside the
image. We used the same shade guide to take 10 different
images per shade tab at the fixed shutter speed (1/200 s), F-
stop (F25), automated white balance and ISO 100. All other
settings of the camera were set to automatic/default mode.
The captured images were outputted in the defaulted jpeg
format with the size of 2592  3888 pixels each. The jpeg
images were converted into the bitmap format for subse-
quent processing.
2.3. Preprocessing
2.3.1. Content segmentation by manual cropping
In each of the images captured, only a portion inside the
shade tab was cropped manually to represent its content
(Fig. 3(a)). The main purpose of the cropping was to prevent
the content from being influenced by light shadow on the
lateral and cervical borders of the tab. The window cropped
was close to the maximum inscribed rectangle of the tab,
named as shade content. The shade contents were the
inputted image of our automatic matching system. The
system was implemented using Matlab (Matlab R2007b,
MathWorks, USA). Following the user-cropping content
segmentation, the rest of processing procedures were carried
out automatically by the system.
the contents were first converted to greyscale images.
The threshold, which was 180 in our experiments, was
trained from the over-reflected regions in the image data.
The pixels in the original color contents whose grey levels
are greater than the threshold, reflecting the high intensity
of the over-exposed regions, were excluded from shade
calculation.
2.3.3. Blocks segmentation
The entire contents were divided into m  n small blocks in
order to compute the color distribution. In our experiments,
we set m = 10 and n = 2. The segmentation was shown in
Fig. 3(b).
2.4. Features extraction
Two device-dependent (RGB, HSV) and two device-independent
color spaces (XYZ, La*b*) were evaluated in this study. The
conversions from RGB to HSV were expressed in Eqs. (1a) to (1f),
conversion to XYZ in Eq. (2). The La*b* color space was converted
from XYZ and expressed in Eqs. (3a) to (3d). The features of the
color spaces were extracted from all blocks and evaluated.
Statistic measurements (mean values) were calculated for each
block to represent its color statistics.
The conversion from RGB to HSV is defined as below:

u if B  G
H¼ ; (1a)
360  u if B > G
8
< 1  m if M > 0
S M ; (1b)
:
0 if M ¼ 0
M
V¼ ; (1c)
255
where

( )
2.3.2. Light reflection removal by thresholding ð1=2Þ½ðR þ GÞ þ ðR  BÞ
The shade contents were subject to noise mainly from over- u ¼ cos1 1=2
; (1d)
½ðR  GÞ2 þ ðR  BÞðG  BÞ
reflection of flashlight on most contour areas. These areas
were removed through a thresholding procedure in order
to prevent from affecting shade features. In this procedure, M ¼ maxfR; G; Bg; (1e)
e6 journal of dentistry 40s (2012) e3–e10
[(Fig._3)TD$IG]

Fig. 3 – The shade distance measurement (a) manual cropping on a shade image; (b) a content after noise removal and
division into smaller blocks and (c) the shade distance between two contents.

m ¼ minfR; G; Bg: (1f) The matching accuracy was defined as an average outcome
rate of which the predicted labels included the same labels of
The conversion from RGB to XZY is through the following the shade tab images being examined, as expressed in Eq. (5),
matrix transformation:
the number of relevant retrieved shade tabs
accuracy ¼ (5)
2 3 2 3 2 3 the number of retrieved shade tabs
X 0:412453 0:357580 0:180423 R
6 7 6 7 6 7
4 Y 5 ¼ 4 0:212671 0:715160 0:072169 5 4G5: (2) For the top n match, the accuracy was counted for any relevant
Z 0:019334 0:119193 0:950227 B shade tabs found within the set of n retrieved shade tabs.

The conversion from XYZ to La*b* is defined as follows:

  3. Results
Y
L ¼ 116  f  16; (3a)
YW
The average size of the cropped contents was 10,346 pixels,
     and the average height and width were 129 and 80 respective-
X Y
a ¼ 500 f  f ; (3b) ly.
XW YW

     3.1. Features selection

Y Z
b ¼ 200 f  f ; (3c)
YW ZW
The comparison between color spaces revealed that L*a*b* and
where HSV had the accuracy 0.75 and 0.67 respectively, higher than
p ffiffiffi
3
q q > 0:008856 the others, RGB (0.55) and XYZ (0.50). Our study focused only on
f ðqÞ ¼ (3d)
7:787q þ 16=116 q  0:008856 the color features extracted from the two impressive color
spaces, L*a*b* and HSV. Individual feature comparison showed
XW, YW and ZW are the given reference white tristimulus that there were higher accuracy of b* feature (0.80) from L*a*b*
values of CIE standard D55 illumination, defined by and S (0.75) in HSV than the others, a* (0.63), V (0.43), H (0.33), L
x = 0.3324 and y = 0.3474 in the CIE chromaticity coordinate. (0.32).

2.5. Similarity measurement 3.2. Performance of proposed metrics

To measure the similarity between two contents, a measure-
ment of the block distances was used in this study. The shade
distance was summation of the distances between two
corresponding blocks in contents s and t. As shown in
Fig. 3(c), it was defined as:
X s 2
distðs; tÞ ¼ ð ðb ðiÞ  bt ðiÞÞ Þ1=2 ;
i
where i was the block label within the content, 1 2 i 2 m  n;
bs and bt were the blocks in contents s and t, respectively. The
top n shade candidates were matched finally when the simi-
larity measurements between the test and the reference shade
contents were the minimum n shades. In other words, the top
one match referred to the results that considered only the first
choice, where n = 1, while the top five matches considered the
first five choices, where n = 5.
The feature S in HSV space, a* and b* in L*a*b* space were the
best three features in shade matching. The performance of
these features and their possible combinations, Sa*b* and Sb*,
were evaluated. The La*b* feature set, which was equivalent to
the DE distance between two contents, was also compared.
The performance of these feature sets were shown in Fig. 4(a).
(4) The authors inferred that the Sa*b* feature set was better than
the others and was regarded as the most appropriate metric in
this study. The a* feature in L*a*b* color space had poor
matching capability by its own; however, it performed better
when used together with S and b* features.
3.3. Reference images
In the following discussion, all-image represented the set of all
captured images, while mean-image the set of 26 mean images
 journal of dentistry 40s (2012) e3–e10 e7
[(Fig._4)TD$IG]
(a) 1
matching probability, we accepted the top five matching
0.9 results in this study. Fig. 4(b) showed the performance of the
0.8 top five matching results of two reference sets using the Sa*b*
0.7 feature sets. According to the tendency of top five matches, the
precision

0.6 accuracy of all-image in top three matches (0.87, 0.94 and 0.94)
0.5 increased obviously and became flattened thereafter. It
mean-image
0.4 reached the highest rates of 0.95 at the top five matches.
all-image
0.3
0.2 3.4. Stability between individual shoots
0.1
0 The shade tab samples being captured at varying time
S a* b* Sa*b* Sb* ΔE
intervals had varying matching accuracies. The accuracies
color features
of individual shoots were ranged from 0.73 to 0.96 as shown in
1 Fig. 5.
(b)
0.95 3.5. The accuracy of shade tabs

0.9 The shade tab accuracies were labelled on the tabs in Fig. 6. The
precision

3D-MASTER shade guide groups the tabs into five categories

0.85
0.8
0.75
0.7
1 2 3 4
number of matches
Fig. 4 – (a) Accuracy using different feature sets. S, a* and b*
representing the color features in HSV and La*b* color
spaces. (b) The performance of two reference sets in top
five matches using Sa*b* feature set.
of the shade tabs calculated from the captured images. The
measurement using all-image contents had a higher accuracy
than that using mean-image contents at the top one matches,
0.87 in all-image and 0.82 in mean-image. In order to have more
options for dentists on shade selection and provide higher
[(Fig._5)TD$IG]
mean-image according to the lightness. Group 1 (on the left of the shade
all-image guide) refers to the lighter and group 5 (on the right) the darker
tab color. The average matching accuracies with respect to
lightness were also depicted in the figure. It trended to have
more accurate matching on darker shades, average accuracy of
0.99 and 0.97 in the lightness groups 4 and 5 respectively.
5
3.6. Result images
Fig. 7 showed top five matched contents using the Sa*b* feature
set. The images of test contents were placed on the left and
that of the predicted contents on the right. The predictions
were aligned from left to right representing the top one to five
contents. Though the test label might not be exactly matched
in the first place, the predicted candidates were very similar to
human eyes and indistinguishable in most time. Attention
should be paid to the difference of the contents’ size in the
figure. In our method, the smallest shade distance could be
attained regardless of the content size when they were
cropped manually from the image.

Fig. 5 – Diagram showing the accuracy of individual shoots. The average accuracy is 0.87 with the standard deviation being
0.07. The variations are mainly due to the open experimental environment.
e8 journal of dentistry 40s (2012) e3–e10
[(Fig._6)TD$IG]

Fig. 6 – The accuracy on each shade tabs and of different lightness groups. The tab accuracies are labelled on the tabs. Upper
row represents the average accuracy in each lightness group.
[(Fig._7)TD$IG]

Fig. 7 – The matching results of Sa*b* feature set using all-image. Three test contents are shown on the left and top five
matched contents on the right aligned in ascending order from left to right. Note that matched contents are not necessary
be in the same size.

uncontrolled environmental factors were not the topics to be

4. Discussions
The main contribution in this study was to realise the
application of digital cameras in dental shade matching.
The method measured the shade distance between the shade
images using content-based color descriptions. The shade
distance compared not only the color features, but also the
texture of the shade. Since different size of the descriptors
could be compared, there was no size restriction of the tooth
surface being measured. Apart from colorimetric or spectro-
photometric analysis commonly used in dentistry, our new
content-based analysis provided a method with promising
accuracy (0.87).
The study considered usual habit of the professionals in
taking images using digital cameras in the clinics. In that way,
the illumination was unstable and the actual shooting
distance was seldom fixed. The variations other than the
discussed in this article. If the illumination and the shooting
distance are fixed, the accuracy of this study would be better.
The reason of using the shade tab images, instead of extracted
teeth or in vivo tooth images, for the experiments was owing
to the demand of ground truth. In vivo shade matching will be
explored after reliable matching methods using digital
cameras are established.
The accuracy of shade matching using digital cameras
under clinical environment has seldom been mentioned. Most
researches focus on measuring L, a* and b* using instrumental
techniques to analyse digital Images.32 Though similar La*b*
characteristics should be found on similar shades, satisfactory
matched shades do not necessarily have the smallest DE value.
Even though DE as a color measurement has long been
accepted in dentistry, it was not the best measurement in our
study. Instead of comparing the La*b* value, this study focused
on the actual accuracy of matching on other color spaces. After
 journal of dentistry 40s (2012) e3–e10
features recomposing, Sa*b* feature set has the best accuracy
in our experiments.
For the color features arranged in the shade guide,
lightness has potential influences on matching accuracy.
The results in this study revealed that the darker shades could
be matched with higher accuracy.
The shade guide used in this study was different from the
previous research, reported 50% accuracy using colorimeter16
and 83.3% using spectrophotometer,37 in the way that they use
either Vita Lumin or Vita Classical shade guides, which have
only 16 shade tabs for matching. It is possible that if 16 tabs
were employed in this study, matching performance would
have improved.
The authors also recommended the top three (accuracy of
0.94) matches were helpful to the dentists for shade
references. Since the shade colors are indistinguishable
enough in some instances, they are not differentiable by
either visual selection or computer inspection. The top three
predicted shade tabs should be placed with the tooth in the
images to facilitate the communication for the fabrication of
porcelain in the laboratory.38,39
For the clinical application in the future, the software
implemented with the proposed matching method needed to
be installed in a computer. A digital image was taken from the
patient’s mouth and read by the software. The program would
provide three to five shade tab suggestions which were the most
similar color to the intra-oral teeth. The computing process will
be finished in several seconds. This information could either be
sent directly to the dental laboratory or be used as the references
for taking the second matching image before sending to the
laboratory. If the second image was required, the suggested
shade tabs should be placed next to the teeth inside the mouth
for real references. In the dental laboratory, the technician could
fabricate the prostheses considering both the suggested shade
tabs and the teeth texture shown on the image.
5. Conclusion
This study proposed a content-based method for dental shade
matching using digital camera. The method reassembled
effective color features and comprised shade gradation and
texture for comparison. The accuracy of the proposed method
was comparable to that of colorimetric or spectrophotometric
analysis. It leaded to a conclusion that digital camera, when
used with the proposed method, might be a tool for dental
shade matching.
Conflict of interest
The author has no conflicts of interest and our article is not
supported by any funding body.
references
1. Sproull RC. Color matching in dentistry. Journal of Prosthetic
Dentistry 1974;31(Pt 3):146–54.
e9
2. Preston JD. Current status of shade selection and color
matching. Quintessence International 1985;16:47–58.
3. Seghi RR, Hewlett ER, Kim J. Visual and instrumental
colorimetric assessment of small color differences on
translucent dental porcelain. Journal of Dental Research
1989;68:1760–4.
4. Billmeyer FW, Saltzman M. Principles of color technology.
2nd ed. New York;: John Wiley & Sons; 1981. p. 1–110.
5. Van der Burgt TP, ten Bosch JJ, Borsboom PC, Kortsmit WJ. A
comparison of new and conventional methods for
quantification of tooth color. Journal of Prosthetic Dentistry
1990;63:155–62.
6. Spencer LM. Shade selection environment and technique.
Dental Clinics of North America 1996;52:358–62.
7. Pizamiglio E. A color selection technique. Journal of Prosthetic
Dentistry 1991;66:592–6.
8. Miller LL. Shade matching. Journal of Esthetic Dentistry
1993;5:143–53.
9. O’Brien WJ, Johnston WM, Fanian F. Double-layer color
effects in porcelain systems. Journal of Dental Research
1985;64:940–3.
10. Ma T, Johnston WM, Koran III A. The color accuracy of
the Kubelka–Munk theory for various colorants in
maxillofacial prosthetic material. Journal of Dental Research
1987;66:1438–44.
11. Rugh EH, Johnston WM, Hesse NS. The relationship between
elastomer opacity, colorimeter beam size, and measured
colorimetric response. International Journal of Prosthodontics
1991;4:569–76.
12. Davis BK, Johnston WM, Saba RF. Kubelka–Munk reflectance
theory applied to porcelain veneer systems using a
colorimeter. International Journal of Prosthodontics 1994;
7:227–33.
13. Johnston WM, Ma T, Kienle BH. Translucency parameter of
colorants for maxillofacial prostheses. International Journal of
Prosthodontics 1995;8:79–86.
14. Johnston WM, Hesse NS, Davis BK, Seghi RR. Analysis of
edge-losses in reflectance measurements of pigmented
maxillofacial elastomer. Journal of Dental Research
1996;75:752–60.
15. Johnston WM, Kao EC. Assessment of appearance match by
visual observation and clinical colorimetry. Journal of Dental
Research 1989;68:819–22.
16. Okubo SR, Kanawati A, Richards MW, Childress S.
Evaluation of visual and instrument shade matching. Journal
of Prosthetic Dentistry 1998;80:642–8.
17. Cho BH, Lim YK, Lee YK. Comparison of the color of natural
teeth measured by a colorimeter and Shade Vision System.
Dental Materials 2007;23:1307–12.
18. Wozniak WT. Proposed guidelines for the acceptance
program for dental shade guides. Chicago: American Dental
Association; 1987. p. 1–2.
19. CIE (Commission Internationale de l’Eclairage).
Recommendations on Uniform Color Spaces. Color-
difference Equations. Psychometric Color Terms.
Supplement No. 2 to CIE Publication No. 15 (E-13.1)1971/(TC-
1.3). Bureau Central de la CIE, Paris; 1978.
20. Haywood VB, Leonard RH, Nelson CF, Brunson WD.
Effectiveness, side effects, and long-term status of
Nightguard Vital bleaching. Journal of the American Dental
Association 1994;125:1219–26.
21. Wee AG, Lindsey DT, Kuo S, Johnston WM. Color accuracy of
commercial digital cameras for use in dentistry. Dental
Materials 2006;22:553–9.
22. Jarad FD, Russell MD, Moss BW. The use of digital imaging
for colour matching and communication in restorative
dentistry. British Dental Journal 2005;199:43–9.
23. Bengel W. The ideal dental photographic system.
Quintessence International 1993;24:251–7.
e10 journal of dentistry 40s (2012) e3–e10
24. Dozic A, Kleverlaan CJ, Aartman IH, Feilzer AJ. Relation in
color of three regions of vital human incisors. Dental
Materials 2004;20:832–8.
25. O’Brien WJ, Hemmendinger H, Boenke KM, Linger JB, Groh
CL. Color distribution of three regions of extracted human
teeth. Dental Materials 1997;13:179–85.
26. Ishikawa NS, Ishibashi K, Tsuruta O, Weber HP.
Reproducibility of tooth color gradation using a computer
color-matching technique applied to ceramic restorations.
Journal of Prosthetic Dentistry 2005;93:129–37.
27. CIE (Commission Internationale de l’Eclairage)
Recommendations on uniform color spaces, color difference
equations, psychometric color terms. Supplement No.2 to
CIE publication No.15 (E.-1.3.1) 1971/(TC-1.3), Bureau Central
de la CIE, 4 Av. du Recteur Poincaré, 75782 Paris Cedex 16,
Paris, France; 1978.
28. Sproull RC. Color matching in dentistry. Journal of Prosthetic
Dentistry 1973;29(Pt 1):416–24.
29. Sproull RC. Color matching in dentistry. Journal of Prosthetic
Dentistry 1973;29(Pt 2):556–66.
30. Munsell AH. A color notation. 11th ed. Baltimore: Munsell
Color Co.; 1961.
31. Color Television Success in Test. New York Times. 1940
August 30; p. 21.
32. CBS Demonstrates Full Color Television. Wall Street Journal.
1940 September 5; p. 1.
33. Television Hearing Set. New York Times. 1940 November 13;
p. 26.
34. Morton, David L. Television Broadcasting. A History of Electronic
Entertainment Since 1945. 1999. IEEE.ISBN 0-7803-9936-6.
35. CIE (Commission Internationale de l’Eclairage). Commission
internationale de l’Eclairage proceedings. Cambridge:
Cambridge University Press; 1931.
36. Corciolani G, Vichi A, Goracci C, Ferrari M. Colour
correspondence of a ceramic system in two different shade
guides. Journal of Dentistry 2009;37:98–101.
37. Paul S, Peter A, Pietrobon N, Hämmerle CHF. Visual and
spectrophotometric shade analysis of human teeth. Journal
of Dental Research 2002;81:578–82.
38. Chu SJ, Trushkowsky RD, Paravina RD. Dental color
matching instruments and systems. Review of clinical and
research aspects. Journal of Dentistry 2010;38s:e2–16.
39. Joiner A, Hopkinson I, Deng Y, Westland S. A review of tooth
colour and whiteness. Journal of Dentistry 2008;36:2–7.