electronics

Article
A Fast Automatic Reconstruction Method for Panoramic Images
Based on Cone Beam Computed Tomography
Jianguo Zhang 1 , Yichuan Jiang 1 , Fei Gao 1 , Sheng Zhao 1 , Fan Yang 1 and Liang Song 2, *

1 School of Mechanical Engineering, Shanghai Institute of Technology, Shanghai 201418, China;

[email protected] (J.Z.); [email protected] (Y.J.); [email protected] (F.G.);
[email protected] (S.Z.); [email protected] (F.Y.)
2 Department of Stomatology, Shanghai Fifth People’s Hospital, Fudan University, Shanghai 200240, China
* Correspondence: [email protected]

Abstract: Panoramic images have been widely used in the diagnosis of dental diseases. In the process
of panoramic image reconstruction, the position of the dental arch curve usually affects the quality
of display content, especially the completion level of the panoramic image. In addition, the metal
implants in the patient’s mouth often lead the contrast of the panoramic image to decrease. This paper
describes a method to automatically synthesize panoramic images from dental cone beam computed
tomography (CBCT) data. The proposed method has two essential features: the first feature is that
the method can detect the dental arch curve through axial maximum intensity projection images over
different ranges, and the second feature is that our method is able to adjust the intensity distribution
of the implant in critical areas, to reduce the impact of the implant on the contrast of the panoramic
image. The proposed method was tested on 50 CBCT datasets; the panoramic images generated
by this method were compared with images attained from three other commonly used approaches
and then subjectively scored by three experienced dentists. In the comprehensive image contrast
score, the method in this paper has the highest score of 11.16 ± 2.64 points. The results show that the
Citation: Zhang, J.; Jiang, Y.; Gao, F.; panoramic images generated by this method have better image contrast.
Zhao, S.; Yang, F.; Song, L. A Fast
Automatic Reconstruction Method Keywords: panoramic image; dental arch detection; image enhancement
for Panoramic Images Based on Cone
Beam Computed Tomography.
Electronics 2022, 11, 2404.
https://doi.org/10.3390/ 1. Introduction
electronics11152404
In the field of modern dentistry, cone beam computed tomography (CBCT) is currently
Academic Editor: Byung Cheol Song a diagnostic imaging technique widely used in assisted diagnosis, virtual simulation,
Received: 14 June 2022
and treatment planning [1,2], with multiple clinical applications in oral implants [3–6].
Accepted: 26 July 2022
The panoramic image displays tissue information, such as the maxillary, mandible, and
Published: 1 August 2022
dentition in one image by projecting the oral CBCT data within a certain range, providing
the dentist with an intuitive image of the oral and maxillofacial tissues [7]; it plays an
Publisher’s Note: MDPI stays neutral
indispensable role in the post-processing of oral images. In addition, research on tooth
with regard to jurisdictional claims in
segmentation based on panoramic images [8], mandibular neural canal segmentation [9],
published maps and institutional affil-
mandibular fracture detection [10], and oral 3D reconstruction [11] has gradually increased.
iations.
Therefore, the automatic panoramic image reconstruction method can provide doctors with
accurate images and provide a data basis for follow-up research.
In general, there are two reconstruction methods for oral panoramic images: the
Copyright: © 2022 by the authors.
first method is to construct the image based on the dental arch line, through the curved
Licensee MDPI, Basel, Switzerland. planar reformation (CPR) [12], which directly generates the corresponding multi-plane
This article is an open access article reconstruction (MPR) image of the curve. This method can quickly reconstruct the image
distributed under the terms and but may lead to the inability to fully display the root of the tooth, abnormal dentition, or
conditions of the Creative Commons implant in the upper and lower jaws [13,14]. The second method is to form the image based
Attribution (CC BY) license (https:// on the curve and thickness of the dental arch. The data within the thickness range of the
creativecommons.org/licenses/by/ dental arch are rearranged to produce the curved multiple MPR image sets, and the final
4.0/). oral panorama image is reconstructed by ray-summation or X-ray method [15,16], which

Electronics 2022, 11, 2404. https://doi.org/10.3390/electronics11152404 https://www.mdpi.com/journal/electronics

Electronics 2022, 11, 2404 2 of 11

displays the internal structure of the oral cavity on a higher completion level. However,
this approach often brings unnecessary areas into the calculation process and raises the
calculation cost, which results in reduced contrast and blurry images. Therefore, finding
a small and suitable arch curve is the key to the reconstruction process.
In order to obtain accurate, high-quality panoramic images [17], related research
has proposed a variety of methods for fitting dental arches. Chanwimaluang et al. [18]
used a polynomial equation to fit the dental arch to generate a panoramic image from the
dental arch curve. After that, more and more studies used different curves and control
points [19,20] to fit the dental arch more accurately. To determine the shape of the dental
arch in different patients, Bae et al. [21] used tooth position-controlled cubic B-splines
to successfully fit different patient arch shapes. In terms of automatic reconstruction of
panoramic images [22–24], Sa-Ing V et al. [25] introduced a dental arch detection algorithm
based on threshold and image morphology algorithms. The dental arch calculation is
simplified by selecting a slice of the original CBCT for downsampling. After the dental
arch is detected, a series of curved MPR image sets are generated to produce a ray-sum
panoramic image. In order to show more of the inside of the mouth in the panoramic
image, Luo T et al. [26] described a method for reconstructing panoramic images based on
three-dimensional surfaces. First, the dental arch curve was generated in axial maximum
intensity projection (MIP), and the long axis curve was generated according to the shape of
the long axis of the upper and lower teeth, extracting a three-dimensional surface through
the curve to describe the entire dentition, and then expanding the three-dimensional
surface to obtain a panoramic image. The panoramic images generated by this method
can make the slices very thin to obtain sharper images and easily recognizable anatomical
structures. In order to automatically obtain a high-contrast panoramic image, Yun Z
et al. [27] introduced an automatic extraction method for dental arch curves. The coronal
MIP histogram is fitted by the normal curve, the range of the transverse MIP section is
determined according to the peak to detect the arch line and the arch thickness area, the
curved MPR image set is generated in the arch thickness range, and finally, the panoramic
image is reconstructed using the synthesis enhancement algorithm. All of these methods
can automate the processing of oral CBCT data to reconstruct the panoramic image, but
there are still problems with the contrast caused by metal implants in the panoramic image.
Based on previous research, we found that in the process of generating the dental
arch curve, the dental arch can be divided into multiple regions according to the growth
of the teeth, and different dental arches can be generated, respectively. To avoid shifting
caused by anomalies and calculation errors, we propose a dental arch curve detection and
adjustment method based on axial MIP images over different ranges. This method displays
the entire dentition with a high level of completion in the panoramic image. On the other
hand, many studies proposed different reconstruction methods to improve the contrast
of panoramic images, while ignoring the contrast problems caused by metal implants in
panoramic images. To resolve contrast issues, the strength of the metal implant is altered
during the image extraction and fusion process to reduce the impact on other tissue areas
in our method. By applying this technique, the contrast and quality of panoramic images
were significantly improved.
The rest of this paper proceeds as follows. The full flow of the panoramic image algo-
rithm is introduced in Section 2. In Section 3, the comparative experiments and evaluation
results of different methods are presented, and the advantages and disadvantages of the
proposed method are discussed. Finally, we make a conclusion in Section 4 and propose
future directions. This method improves the contrast of panoramic images by altering the
strength of the metal implant to reduce the impact on other tissue areas.

2. Materials and Methods

2.1. Data Acquisitions
The 50 clinical dental CBCT samples employed in this paper were collected from ran-
domly selected patients by the Shanghai Fifth People’s Hospital and Zhongshan Hospital,
 2. Materials and Methods
2. 2.1.
Materials and Methods
Data Acquisitions
2.1. Data
TheAcquisitions
50 clinical dental CBCT samples employed in this paper were collected from ran-
Electronics 2022, 11, 2404 domlyThe selected
50 clinical dental by
patients CBCT samples employed
the Shanghai in this
Fifth People’s paper were
Hospital collected from
and Zhongshan 3 of 11
ran-
Hospital,
domly
Fudan selected patients
University, by the Shanghai
Shanghai, China, and FifththePeople’s
involved Hospital and Zhongshan
equipment was acquired Hospital,
using a
Fudan
PlanmecaUniversity,
ProMaxShanghai, China, andFinland)
3D Max (Planmeca, the involved
scanner equipment
and NewTom was acquired
VGi (QR s. using a
r. l, Ve-
Planmeca
rona, ProMax
Italy) scanner3D Max
with (Planmeca,
the following Finland) scanner
parameters: 110 and
kVp,NewTom
Fudan University, Shanghai, China, and the involved equipment was acquired using a3–8 mA, VGi
scan (QR s.
time: r.
27l, Ve-
s. The
rona,
size Italy)
of all scanner
Planmeca slices
ProMax ofwith
3D the
CBCT Max following
was parameters:
776 × 776
(Planmeca, pixels,
Finland) 110
with akVp, 3–8
0.2mA,
0.2 ×and
scanner mm scan
NewTompixeltime:
size.27
VGi The
(QR s. s.
The
inter-
r. l,
size of distance
slice
Verona,all slices of
Italy)was CBCT
0.2 mm.
scanner was
withThe 776
the × 776 pixels,
number of slices
following with
wasaaround
parameters: 0.2110
× 0.2 mm
432,
kVp, pixel
depending
3–8 mA, size.
onThe
scan the inter-
patient
time: 27 s.
slice distance
size of was
situation.
The 0.2 mm.
all slices The number
of CBCT was 776 of ×
slices
776was around
pixels, with432, depending
a 0.2 × 0.2 mmon the patient
pixel size. The
situation.
inter-slice distance was 0.2 mm. The number of slices was around 432, depending on the
2.2. Methods
patient situation.
2.2. Methods
As shown in Figure 1, the workflow of our proposed method was mainly divided
2.2. Methods
intoAs twoshown
steps.inFirst,
Figure
the 1, theMIP
axial workflow
image ofof the
ourmandible
proposedand method wasMIP
the axial mainly
image divided
without
into two As shown
steps. in
First, Figure
the 1,
axial the
MIP workflow
image of of
the our proposed
mandible and method
the
the crown were calculated according to the coronal MIP, and the dental arch curve was was
axial MIPmainly
image divided
without into
thetwo steps.
crown were
detected; First, the
then,calculatedaxial
the enhanced MIP
according image
image to of the mandible
theextracted
was coronal MIP, and
from and the axial
the dental
the curved MIP
MPR archimage
curve
image without
set,was
and
thefinal
detected;
the crownthen,were
the calculated
panoramic enhanced
image was according
image wasto
obtained the coronal
extracted
using thefrom MIP,
image and
thefusion
curvedthe dental
MPR
algorithm. arch curve
image set, and was
detected; then, the enhanced image was extracted from
the final panoramic image was obtained using the image fusion algorithm. the curved MPR image set, and the
final panoramic image was obtained using the image fusion algorithm.
Axial MIP
(Dental crown)
Axial MIP Largest
Coronal MIP (Dental crown) connected
CBCT Coronal MIP Largest
mask component
Coronal MIP connected
CBCT Coronal MIP Axial MIP
mask component
(Without dental
Axial MIP
crown)
(Without dental
crown)
Obtain
Panoramic Enhanced Curved MPR
reference
image image images Obtain
Panoramic Enhanced Curved MPR curve
reference
image image images
curve

Figure
Figure1.1.Workflow
Workflow of
of the
the proposed
proposed method.
method.
Figure 1. Workflow of the proposed method.
2.2.1.Dental
2.2.1. DentalArchArchCurve
CurveDetection
Detection
2.2.1. Dental
At Arch
Atpresent, Curve
present, the
the main Detection
main dental
dental arch
arch detection
detectionmethod
methodisistotodetect
detectthe
thedental
dentalarch
archregarding
regard-
theAt
ing axial
the MIP.MIP.
present,
axial If
thethe Ifaxial
main MIP
axialisarch
thedental directly
MIP reconstructed
detection
is directly method using
is
reconstructed CBCT
to detect
using data,
the excessive
dental
CBCT amounts
arch excessive
data, regard-
ingofthe
irrelevant
amounts axial tissue
ofMIP. willtissue
If the
irrelevant be displayed
axial MIP
will is in the axial
bedirectly MIPthe(Figure
reconstructed
displayed in 2a), which
axialusing
MIP (Figurewill
CBCT 2a),result
data, in poor
excessive
which will
accuracy
amounts of
result inofpoor the dental
irrelevant
accuracy arch
tissue detection.
will
of the If a
be displayed
dental certain range
in theIfaxial
arch detection. of data is needed
MIP range
a certain (Figure to
of 2a),reconstruct
datawhich willthe
is needed to
axialinMIP,
result poorthe
reconstruct the range
accuracy isofnormally
axial MIP, the
thedentalpreset
range arch and manually
detection.
is normally modified
If a and
preset certain range
manuallyby the
of user,iswhich
data
modified by thecauses
needed to
user,
undesired
reconstruct
which interactions
the
causes axial MIP, the
undesired that cannot
range
interactions be
thatautomated.
is normally preset
cannot The MIP images
and manually
be automated. generated
Themodified
MIP images from
by the four
user,
generated
different
which
fromcauses ranges of
undesired
four different data are shown
interactions
ranges in Figure
thatshown
of data are 2.
cannotinbeFigure
automated.
2. The MIP images generated
from four different ranges of data are shown in Figure 2.

(a) (b) (c) (d)

(a) (b) (c) (d)
Figure
Figure2.2.MIP
MIPimage
imagegeneration
generationusing different
using data
different ranges.
data (a) Complete
ranges. original
(a) Complete CBCT
original data.data.
CBCT (b)
Coronal
(b) Coronal MIP includes the upper and lower jaw. (c) Coronal MIP of mandible. (d) Coronal MIPin
Figure 2. MIP
MIP includes
image the upper
generation and
using lower jaw.
different (c)
data Coronal
ranges. (a)MIP of mandible.
Complete (d)
original Coronal
CBCT MIP
data. (b) in
which MIP
Coronal the mandible
includes does
the not contain a crown.
which the mandible doesupper and lower
not contain jaw. (c) Coronal MIP of mandible. (d) Coronal MIP in
a crown.
which the mandible does not contain a crown.
In
Inthe
theprocess
processof ofgenerating
generatingMIP
MIPimages,
images,the
theidea
ideaisisto
toencode
encodethe
thevolume
volumedata
dataof
ofthe
the
In
largestthe process
signal on of
the generating
trajectory MIP images,
projected, the
along idea
with is
the to encode the
perspective ofvolume
the data
coronal of the
surface,
largest signal on the trajectory projected, along with the perspective of the coronal surface,
largest signal on
and project thethe trajectory projected,
high-intensity signal ofalong with the perspective
the three-dimensional of the
space intocoronal
a planesurface,
to form a
continuous image.
In the coronary MIP image, the data on enamel, cortical bone, and soft tissue were the
most numerous, and the Hu values of these three tissues were quite different, so they were
represented as three independent peaks in the Hu intensity histogram, of which the largest
peak at the abscissa was the distribution of the Hu value of the tooth. The tissue Hu value
 and project the high-intensity signal of the three-dimensional space into a plane to form a
continuous image.
In the coronary MIP image, the data on enamel, cortical bone, and soft tissue were
the most numerous, and the Hu values of these three tissues were quite different, so they
Electronics 2022, 11, 2404
were represented as three independent peaks in the Hu intensity histogram, of which the
4 of 11
largest peak at the abscissa was the distribution of the Hu value of the tooth. The tissue
Hu value in the histogram can be obtained by fitting Formula (1). First, the generated Hu
intensity histogram is smoothed using Gaussian smoothing to remove the violent fluctu-
in the histogram
ations in the localcan be obtained
areas by fittingand
of the histogram Formula
reduce(1).
theFirst,
impactthe on
generated HuThen,
the crests. intensity
the
histogram is smoothed using Gaussian smoothing to remove the violent fluctuations
histogram curve is derived to satisfy Formula (2), where the first derivative is 0 and in the
the
local areas of the histogram and reduce the impact on the crests. Then, the histogram curve
second derivative is negative to determine the peak value and N. Finally, the final curve
is derived to satisfy Formula (2), where the first derivative is 0 and the second derivative is
is obtained by fitting a Gaussian multimodal function [28].
negative to determine the peak value and N. Finally, the final curve is obtained by fitting
𝑁 (𝑥−𝜇𝑖 )2
a Gaussian multimodal function [28]. 𝐴 − 𝑖
2
2𝛿𝑖
𝑓 (𝑥 ) = ∑ 𝑒 (1)
( x − µ i )2
N𝑖=1
√2𝜋𝛿
Ai 2
−
∑q
2
fx = 𝑖
e 2δi (1)
i =1 2πδi2 2
𝑑𝑝 𝑑 𝑝
= 0, <0 (2)
𝑑𝑥 2
dp d2 p𝑑𝑥
= 0, <0 (2)
where 𝐴𝑖 is the peak gain corresponding dx dx2 peak, and the sum is equal to one; 𝜇𝑖 is
to each
the mean
where Ai isvalue corresponding
the peak to each peak;
gain corresponding to each 𝛿𝑖 peak,
is theand
standard
the sum deviation
is equal to corresponding
one; µi is the
to eachvalue
mean 𝑁 is the number
peak.corresponding of peaks
to each peak;ofδithe Gaussian
is the standardfunction;
deviation 𝑝 is the histogramtocurve.
corresponding each
peak.After
N is fitting the histogram
the number of peaks with
of thea Gaussian function;
multimodal p isfunction, the threshold
the histogram curve. of the
filtered image
After wasthe
fitting obtained
histogramat thewith
peaka with the largest
Gaussian abscissafunction,
multimodal (Figure 3b). theItthreshold
can be seen of
from the coronal
the filtered imageMIPwasimage
obtainedthatatthe
theX-ray
peakabsorption of tooth
with the largest enamel
abscissa is higher
(Figure 3b). than
It canthat
be
of
seenthefrom
uppertheand lowerMIP
coronal jawimage
cortical bones,
that and the
the X-ray intersection
absorption of teeth
of tooth and alveolar
enamel is higherbones
than
can
that be more
of the clearly
upper anddistinguished.
lower jaw corticalThe distribution
bones, and the of intersection
the teeth canofbe obtained
teeth after a
and alveolar
horizontal
bones can be projection of thedistinguished.
more clearly tooth by threshold The filtering (Figure
distribution 3c),teeth
of the with can
the largest peak
be obtained
after
at thea patient’s
horizontaltooth
projection of the
bite near thetooth
upperby and
threshold
lowerfiltering (Figure 3c),
jaw boundaries. Thewith the largest
slicing range
peak
values at of
thecrests
patient’s
andtooth bite near
troughs, both the
oralupper and and
occlusal lower jaw boundaries.
intersections The and
of teeth slicing range
alveolar
values
bone, are of determined
crests and troughs,
by the sameboth process
oral occlusal and above.
described intersections
Then, theof teeth
axialand
MIPalveolar
images
bone,
are are determined
generated from twoby the same
ranges process
of slices. Thisdescribed
procedureabove. Then,the
of splitting theoriginal
axial MIP images
image into
are generated
two fragmentsfrom for two
betterranges of slices.
imaging This procedure
is named sub-regional of splitting
detection, theasoriginal
shownimage into
in Figure
two fragments for better imaging is named sub-regional detection, as shown in Figure 2c,d.
2c,d.

(a) (b) (c)

Figure 3. The
Theaxial
axialslice
slicerange
rangedetection
detectionprocess. (a)(a)
process. Coronal MIP
Coronal images.
MIP (b) Hu
images. (b) histogram of cor-
Hu histogram of
onal MIP.
coronal (c) (c)
MIP. Dental mask
Dental image.
mask image.

The calculation
calculation ofofthe
thedental
dentalarch
archcurve
curveisisestablished
establishedonon the axial
the MIP.
axial MIP.Bone
Bonethreshold-
thresh-
ing parameters
olding parameters are are
calculated using
calculated thethe
using samesamemethod
method andandprovide
provide a filtered mask
a filtered maskimage.
im-
In theInmask
age. image,
the mask the contour
image, formed
the contour by thebyjaw
formed theand
jawteeth is the is
and teeth largest in area,
the largest in and
area,small
and
holes are
small commonly
holes observed
are commonly in the contour.
observed Furthermore,
in the contour. visible burrs
Furthermore, visibleare oftenare
burrs seen on
often
the boundaries
seen of contour,
on the boundaries of which
contour, leads to rough
which leads edges.
to rough Therefore, the contour
edges. Therefore, thepattern
contour is
treated by the expansion corrosion method to obtain a relatively full and
pattern is treated by the expansion corrosion method to obtain a relatively full and smooth smooth contour
(Figure
contour4a). The set
(Figure of The
4a). contour
set points of all points
of contour adjacentofareas is extracted
all adjacent in the
areas smooth contour
is extracted in the
map. Eventually, the contour with the largest area in the image
smooth contour map. Eventually, the contour with the largest area in the is extracted, andimage
the contour
is ex-
image ofand
tracted, the the
mandible
contourand teethofisthe
image obtained.
mandible and teeth is obtained.
 Electronics
Electronics
Electronics 2022,
2022,
2022, 11,
11,
11, 2404
xxFOR
FORPEER
PEERREVIEW
REVIEW 5 of
55 ofof 1211
12

(a)
(a) (b)
(b) (c)
(c) (d)
(d)

(e)
(e) (f)
(f) (g)
(g) (h)
(h)
Figure
Figure
Figure 4.4.4.
Extraction
Extraction
Extraction and
andand transformation
transformation
transformation process
process
process ofofofthe
the
the dental
dental
dental arch
arch
arch curve.
curve.
curve. (a)
(a)(a)Contour
Contour
Contour image
image
image after
after
after
expansion
expansion corrosion.
corrosion. (b)
(b) The
The skeleton
skeleton
expansion corrosion. (b) The skeleton line line
line after
after thinning. (c) Skeleton line after de-bifurcation. (d) (d)
thinning.
thinning. (c)
(c) Skeleton
Skeleton line
line after
after de-bifurcation.
de-bifurcation. (d)
The
The
The dental arch line is fitted by cubic spline
dental arch line is fitted by cubic spline curve. (e)(e)
dental arch line is fitted by cubic spline curve.
curve. (e)The
The The skeleton
skeleton
skeleton line
lineline ofofcoronal
coronal
of coronal MIP MIP
MIP ofofmandible
mandible
of mandible after
after
after de-bifurcation.
de-bifurcation.
de-bifurcation. (f)The
(f)
(f) The The skeleton
skeleton
skeleton line
lineline ofofcoronal
of coronalcoronal
MIPMIP MIP after
afterafter decapitation.
decapitation.
decapitation. (g)Calculation
(g) Calculation
(g) Calculation ofofcurve
of the the
the
curvedistance
curve distanceimage
imageinin(e,f).
(e,f).(h)
(h)The
Thedental
dentalarch
archcurve
curveafter
afterequidistant
equidistanttransformation.
transformation.
distance image in (e,f). (h) The dental arch curve after equidistant transformation.

Detection
Detection
Detection ofof
of the
the the dental
dental
dental arch
arch
arch curve
curve based
curvebased
basedon on mask
onmask images.
maskimages. Firstly,
images.Firstly, Firstly, the
the
the mask
mask
mask image
image
image
(Figure
(Figure 4a)
4a) is
is processed
processed byby the
the thinning
thinning algorithm
algorithm [29–31],
[29–31],
(Figure 4a) is processed by the thinning algorithm [29–31], and the bifurcation endpoint and
and the
the bifurcation
bifurcation endpoint
endpoint
template
template
template ofof
of 88fields
8fields
fields isisisadded
added
addedto totocompare
compare
comparethe the skeleton
theskeleton
skeleton line
line
line ininthe
in the refinement
therefinement calculation
refinementcalculation
calculation
process
process (Figure
process(Figure
(Figure4b), 4b),
4b),the the bifurcation
thebifurcation and
bifurcation and endpointendpoint structure
endpoint structure
structureininthein the skeleton
theskeleton
skeletonare are
are recorded,
recorded,
recorded, and
and
and the
thethe branching
branching
branching statistics
statistics
statistics andand
and fork
fork
fork removal
removal
removal are
areare performed
performed
performed (Figure
(Figure
(Figure 4c).
4c).
4c). Then,
Then,
Then, the the
the inser-
inser-
insertion
tion
tion point
point
point isisdetermined
determined
is determined on on onthe
the the skeleton.
skeleton.
skeleton. AArecent
A recentrecent study
study
study [32]
[32][32]
shows shows
shows that
thatthat aasmooth
a smoothsmooth dental
dental
dental arch
arch
arch can
can be
be fitted
fitted by
by 1111 insertion
insertion points
points with
with sufficient
sufficient fitting
fitting
can be fitted by 11 insertion points with sufficient fitting accuracy and low computational accuracy
accuracy and
and low
low computa-
computa-
tional
tional cost.
cost.cost.
Through Through
Through conducted
conducted
conducted experiments,
experiments,
experiments, 13points
13
13 points points were
werewere
usedusedused to
to satisfytosatisfy
satisfy thefitting
the
the fitting fitting ac-
ac-
accuracy
curacy
curacy requirement.
requirement.
requirement. The insertion
The insertion
The insertion points points
points are determined
are determined
are determined at equal
at equal
at equal intervals intervals
intervals
along along
along the hor-
the hor-
the horizontal
izontal
izontal axis,generating
axis,
axis, generating generating
a non-uniformaanon-uniform
non-uniform cubicB-spline
cubic
cubic B-spline B-spline curveto
curve tocurve
represent torepresent
represent
the dentalthe the dental
dental
arch (Figure arch
arch4d).
(Figure
(Figure
Finally,4d).4d). Finally,
Finally,
in order in
to make inorder
order tomake
to makepoint
the insertion theinsertion
the insertion point
point
fit the dental fitthe
fit
arch thedental
more dental archmore
arch
accurately, more
we accu-
accu-
proposed
rately,
rately, weweproposed
an interpolationproposed point
ananinterpolation
interpolation
calibration pointcalibration
method.
point calibration
As shownmethod.method.
by the yellow
AsAsshown
shown
lines bybythe
and the
red yellow
dots in
yellow
lines
Figure and 5,red
at dots
the in Figure
insertion 5,
pointat the
withininsertion
the rangepoint
of within
teeth, a the range
calibration
lines and red dots in Figure 5, at the insertion point within the range of teeth, a calibration of teeth,
line is a
madecalibration
with the
line
line isismade
occlusalmade with
slice
with theocclusal
value,
the occlusal
and thesliceslice value,
position
value, and
ofand the
thethe position
insertion
position ofthe
point
of the insertionpoint
isinsertion
adjusted point
by isisadjusted
adjusted
calibrating the
by
by calibrating
tooth
calibrating thetooth
center position
the tooth center
oncenter
the line.position
The method
position ononthethe line.
can
line. Themethod
reduce
The method
the lack of can
can reduce
fitting
reduce thelack
accuracy
the lack of
caused
of
by the
fitting
fitting abnormal
accuracy
accuracy causedcontour
caused bybytheand
the the error
abnormal
abnormal of the and
contour
contour refinement
andthe
theerror
error algorithm
ofofthe and can algorithm
therefinement
refinement display
algorithm the
andstructure
and candisplay
can of thethe
display tooth
the and the
structure
structure ofimplant
of thetooth
the more
tooth andcompletely.
and theimplant
the implantmore morecompletely.
completely.

(a)
(a) (b)
(b) (c)
(c)
Figure
Figure
Figure5.5.5.
Images
Images ofoftooth
Images tooth
of sections
sections
tooth inindifferent
sections different
in locations.
locations.
different (a)Incisor
(a)
locations. Incisor section.
section.
(a) Incisor (b)Canine
(b)
section. Canine section.
section.
(b) Canine (c)
(c)
section.
Molar
Molar section.
section.
(c) Molar section.

Ascan
As canbe
beseen
seenfrom
fromFigure
Figure5,5,teeth
teethin
indifferent
differentpositions
positionshave
havedifferent
differentdegrees
degreesof
of
inclination in space. For example, the molar area (Figure 5c) is nearly vertical; the crown
inclination in space. For example, the molar area (Figure 5c) is nearly vertical; the crown
Electronics 2022, 11, 2404 6 of 11

As can be seen from Figure 5, teeth in different positions have different degrees of
inclination in space. For example, the molar area (Figure 5c) is nearly vertical; the crown
in the cuspid area (Figure 5b) is slightly tilted forward, and the root is slightly backward;
the teeth are significantly tilted in the incisor area (Figure 5c). If the curved MPR image
is reconstructed based on a single arch curve, the internal condition of the tooth in the
incision area is difficult to visualize, and if the curved MPR image is reconstructed based
on the arch line and arch thickness, it is easy to add other non-target areas to participate in
the reconstruction. We propose to reconstruct a curved MPR image set based on multiple
arch curves. The distance of the dental arch curve transformation was calculated based on
the skeletal lines (Figure 4c,f,g) of the axial MIP image of the mandible (Figure 2b) and the
uncrowned axial MIP image (Figure 2c). The size of the curved MPR image set is determined
by the length of the dental arch curve and the thickness of the CBCT data, respectively,
so the curve needs to be isometrically transformed to generate a curved MPR image of
consistent size and fused. The dental arch curve of the new position (Figure 4h) is obtained
after the isometric transformation. Finally, using multiple arch curves to generate a curved
MPR can show more roots or implants without adding too much computational cost.

2.2.2. Image Enhancement and Synthesis Algorithms

Enhanced images were generated from curved MPR image sets. In order to suppress
non-interested tissues and regions, the parts of interest in curved MPR at different angles
were extracted from each panoramic image, and their intensity distribution was amplified.
We propose a nonlinear extraction algorithm. The enhanced image P0 can be generated by
the following formula:
  
N N
αY ∗ ln ∑n=1 e Pn (i,j)/Y , ∑n=1 Pn (i, j) > Y

P0 =   (3)
N N

 ln ∑n=1 e Pn (i,j)/Y , ∑n=1 Pn (i, j) < Y

where the parameter α is the intensity coefficient of the enhanced image. The value of α is
set to 1.35. The parameter Y is the threshold to the Hu of the soft tissue. Pn(i, j) is the i row
j column value of the nth image in the curved MPR image set. N is the number of images.

β ∗ G ( P1 ), P0 < S
P= (4)
γ ∗ G ( P0 ) + (1 − γ) ∗ W ( P0 ), P0 > S

where the parameter β is an adjustment factor that controls the contrast of the non-area
of interest, β = 1.3. The parameter S is the area of interest threshold in the enhanced
image, S = 105. The parameter γ is the weight factor used to control the enhanced details,
γ = 0.4. W represents the bilateral filter function. G represents the Gaussian filter function.
After the generation of P0 , the region of interest in the enhanced image is mapped to
the panoramic image in the new intensity interval by Formula (4) to improve the contrast of
different regions of the image. The pseudo-code of the image enhancement and synthesis
algorithm is shown in Algorithm 1. Lines 1–8 and lines 9–14 in Figure 6 represent Formula
(4) and Formula (5), respectively.
Electronics 2022, 11, x FOR PEER REVIEW 7 of 12

Electronics 2022, 11, 2404 7 of 11

Algorithm 1 Image synthesis and enhancement algorithms

Input: MPRSets(Psets)
Algorithm 1 Image synthesis and enhancement algorithms
Output: P
1: N←Num(Psets)sets )
Input: MPRSets(P
for PnP(i, j) ∈ Pn do
2:Output:
1: N ←Num(Psets )
3: if Pn=1→N(i,j) > Y then
2: for Pn (i, j) ∈ Pn do
4:3: Pif0(i,j) Pn=1→N(i,j)/Y)
Pn=1=→αYln(e
N (i,j) > Y then
5:4: else P0 (i,j) = αYln(ePn=1 →N(i,j)/Y )
6:5: (i,j) = ln(ePn=1→N(i,j)/Y)
P0else
Pn=1 →N(i,j)/Y )
7:6: end ifP0 (i,j) = ln(e
7:
8: end for end if
8: end for
9: if P0(i,j) > Y then
9: if P0 (i,j) > Y then
10:
10: P(i,j)
P(i,j)= =βG(P
βG(P1)1 )
11: else
11: else
12:
12: P(i,j) = γG(P
P(i,j) = γG(P0)0+ ) +(1-γ)W(P
(1-γ)W(P0)0 )
13:end
13: endifif
14:return
14: return P P

(a) (b) (c)

(d) (e) (f)

Figure 6.
Figure 6. Dental
Dental arch
arch detection
detection effect
effect diagram.
diagram. (a–c)
(a–c) Panoramic
Panoramic image
image reconstructed
reconstructed based
based on
on the
the
dental arch curve generated by software. (d–f) Panoramic image reconstructed based on the dental
dental arch curve generated by software. (d–f) Panoramic image reconstructed based on the dental
arch curve
arch curve generated
generated by
by our
our method.
method.

3. Results and Discussion

3.1. Experimental Environment
3.1. Experimental Environment and
and Implementation
Implementation Time
Time
The
The experiments
experimentsininthis paper
this were
paper all performed
were on a computer
all performed with Intel(R)
on a computer with Xeon(R)
Intel(R)
E-2136 [email protected] GH (Dell, Round Rock, TX, USA) and 32GB RAM
Xeon(R) E-2136 [email protected] GH (Dell, Round Rock, TX, USA) and 32GB RAM (Kingston, (Kingston, Fountain
Valley,
FountainCA, USA),CA,
Valley, using Python
USA), 3.8.5
using (G. van
Python Rossum.,
3.8.5 (G. vanAmsterdam, The Netherlands).
Rossum., Amsterdam, We
The Neth-
performed the image reconstruction experiment and recorded the time after
erlands). We performed the image reconstruction experiment and recorded the time after the CBCT data
reading
the CBCT was completed.
data The average
reading was completed.timeThe
spent on thetime
average method
spent based on the
on the reconstruction
method based on
of the dental arch area was 4.58 s. The method proposed in this
the reconstruction of the dental arch area was 4.58 s. The method proposed in paper has a maximum
this paper
time
has aof 5 s and an
maximum average
time of 5 s calculation
and an averagetimecalculation
of 3.04 s. This
timeshows
of 3.04that the shows
s. This reconstruction
that the
method based on
reconstruction multi-arch
method basedcurves can effectively
on multi-arch curvesreduce the computational
can effectively reduce thecomplexity
computa-
and improve
tional efficiency.
complexity and improve efficiency.
3.2. Dental Arch Detection Effect
3.2. Dental Arch Detection Effect
There are several 3D imaging software currently available for reconstructing panoramic
There are several 3D imaging software currently available for reconstructing pano-
images, such as SimPlant (Materialise Inc. Rotherham, UK) and eXam Vision (KaVo Dental
GmbH,images,
ramic such
Biberach, as SimPlant
Germany). (Materialise
In eXam Vision,Inc.
the Rotherham, UK) and
software supports theeXam Visioncreation
automatic (KaVo
Dental
of dentalGmbH, Biberach,
arch curves Germany).
or manual In eXam
creation Vision,arch
of dental the software supports
curves, and the automatic
a panoramic image
creation of dental arch curves or manual creation of dental arch curves,
corresponding to the curve can be obtained. Therefore, we used the same CBCT and a panoramic
data to
image corresponding
compare the panoramic toimages
the curve can be obtained.
corresponding to theTherefore,
dental archwe usedgenerated
curves the same by CBCT
the
data to compare the panoramic images corresponding to the dental arch
eXam Vision software and our proposed dental arch detection method. It can be seencurves generated
by Figure
in the eXam6a–cVision software
that the and our
panoramic proposed
image dental arch
reconstructed detection
by the softwaremethod.
failed Ittocan be
show
seen in
some Figureand
incisors 6a–cmolars.
that theFigure
panoramic image
6d–f are reconstructed
panoramic imagesbygenerated
the software failed
by the to show
dental arch
detection method in this paper, and the pictures show the entire dentition more completely.
 Electronics 2022, 11, x FOR PEER REVIEW 8 of 12

Electronics 2022, 11, 2404 8 of 11

some incisors and molars. Figure 6d–f are panoramic images generated by the dental arch
detection method in this paper, and the pictures show the entire dentition more com-
pletely.
We We that
believe believe
thethat
keythe key for
factor factor for improving
improving the accuracy
the accuracy of theofproposed
the proposed dental
dental arch
detection method over different ranges regarding the performance of dental arch arch
arch detection method over different ranges regarding the performance of dental curve
curve fitting
fitting is to sub-regional
is to apply apply sub-regional detection
detection for reducing
for reducing disturbance
disturbance amongamong various
various ranges
ranges and add a control point adjustment panel for decreasing
and add a control point adjustment panel for decreasing fitting error. fitting error.

3.3. Different Types of Panoramic Image Effects

3.3.
In addition, when the patient
In patient has has implants
implantssuch suchasasdental
dentalcrowns
crownsininthe themouth,
mouth,metalmetal
implants can compress the brightness
implants brightness distribution
distributionof ofother
othertissues
tissueson onthe theimage,
image,resulting
resulting
in an
in anoverall
overalldarkening
darkeningofofthe thefinal
finalpanoramic
panoramicimage. image.This
Thisisisdueduetoto the
the strong
strong ability
ability of of
the
the metal
metal to absorb
to absorb X-rays
X-rays andand the the
weak weak absorption
absorption ability
ability among
among tissues,
tissues, so so
wewe increased
increased the
the strength
strength of tissue
of the the tissueareasareas
outsideoutside of metal
of the the metal species
species plants,
plants, redistributing
redistributing themthem to
to new
new areas,
areas, and we andtook
we took the enhanced
the enhanced imageimage
thatthat improves
improves the the clarity
clarity of theof the boundary
boundary as as
the
the foreground
foreground andoriginal
and the the original
panoramicpanoramic
imageimage after Gaussian
after Gaussian blurring blurring as the back-to
as the background
ground to
increase theincrease
depth and the depth
spatialand spatial
contrast ofcontrast
the finalofpanoramic
the final panoramic
image. image.
The images
The images of ofexperimental
experimentalresults resultsforfor
common
common oraloral
CBCT CBCTdatadataare presented
are presentedin Fig-in
ure 7. 7.
Figure These
These CBCT
CBCT data
data included
includednormal normaldentition,
dentition,missing
missingteeth, teeth,restorations
restorationswith with
crowns, implants, and mixed
crowns, mixed cases,cases, which
whichcovered
coveredmostmosttypes
typesofofdental
dentalimplant
implantcasescasesinin
different locations.
different locations. In these panoramic
panoramic images,images,the thetooth
tootharea
areaandandthe thetreatment
treatmentarea areahave
have
better luminance
better luminance performance.
performance. Figure Figure7a 7ashow
showaapanoramic
panoramicimage imageofofnormal normaloraloraldata
data
generated with good contrast.
generated contrast. Panoramic
Panoramicimages imagesgenerated
generatedfrom fromCBCT CBCTdata dataofofimplant
implant
and filler treatments
and treatmentsare areshown
shownininFigure Figure7b–f,i, thethe
7b–f,i, orientation
orientation andand depthdepthinformation
information of
thethe
of maxillary
maxillary andand maxillary
maxillary implants
implants cancanbebe observed,
observed, andand thethetreatment
treatmentarea areacancanbebe
clearly distinguished
clearly distinguished from from the the surrounding
surroundingtissue tissuewith
withclear
clearshapes,
shapes,and anddistinct
distinctlayers
layers
can be
can be seen
seen in the image. Panoramic images images of of crown
crown restoration
restoration are are shown
shownin inFigure
Figure7e– 7e–i.
i. The
The panoramicimage
panoramic imagedoes doesnot notappear
appearblackblack due
due to the high
high brightness
brightnessofofthe therestored
restored
crown, maintaining
crown, maintainingaabalancedbalancedcontrastcontrastbetween
betweenthe thetreatment
treatmentarea, area,the the tooth
tooth area,
area, andandthe
the surrounding
surrounding tissue,
tissue, and does
and does not require
not require the userthetouser to adjust
adjust the windowthe windowwidth and width and
window
window
level level tothe
to observe observe the oral
oral tissue tissue information.
information. It can be seenIt canthatbeourseen that ourmethod
proposed proposed has
method
better has better and
brightness brightness
contrast and contrast performance
performance in the tooth in the
areatooth
andareatheand the treatment
treatment area in
area in various
various treatment treatment
situations. situations.

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

Figure 7.
Figure 7. Different
Differentpanoramic
panoramicimage
imageeffects.
effects.(a)
(a)Panoramic
Panoramicimage
imageofofnormal
normaloral.
oral. (b,c)
(b,c) Panoramic
Panoramic im-
image
age of implant
of implant in molar
in molar area. area. (d) Panoramic
(d) Panoramic image image of the implant
of the implant in thearea.
in the canine canine
(e–g)area. (e–g)
Panoramic
image of crown restoration in molar area. (f–i) Panoramic image of crown restoration in canine area.
Electronics 2022, 11, x FOR PEER REVIEW 9 of 12

Panoramic image of crown restoration in molar area. (f–i) Panoramic image of crown restoration in
canine area.
Electronics 2022, 11, 2404 9 of 11

3.4. Quality Evaluation of the Panoramic Radiographs

In this experiment, we divided the image into three regions: irrelevant tissue region,
3.4. Quality Evaluation of the Panoramic Radiographs
tooth region, and treatment region. By referring to the quality requirements of CBCT im-
In this experiment,
ages [27,33–35], we divided
the evaluation contentthe
of image into three
panoramic regions:
images irrelevant
is determined tissue1).region,
(Table The
tooth region,
evaluation andwere
results treatment
scored region.
by threeBy referring
dentists withtoextensive
the quality requirements
clinical experienceofbased
CBCT
onimages [27,33–35],
the evaluation the evaluation
content, content
with a score of panoramic
ranging from 1 toimages is determined
4 points, with a full(Table
score 1).
of The
12
points. An average score was calculated to represent the contrast of each panorama image.on
evaluation results were scored by three dentists with extensive clinical experience based
the evaluation content, with a score ranging from 1 to 4 points, with a full score of 12 points.
Tableaverage
An scoreimage
1. Panoramic was calculated to represent the contrast of each panorama image.
score descriptions.

Score Evaluation
Table 1. Panoramic image score Content
descriptions.
The differences in brightness and contrast between different regions are too large, and the structures
1
within each region areScore
not visible. Evaluation Content
The contrast difference between The differences
different in brightness
regions and contrast
is not obvious, and between differentof
the boundaries regions are too large,
each region are
2 1
and the structures within each region are not visible.
not obvious.
The contrast difference between different regions is not obvious, and the boundaries
3 Brightness and contrast2between ofdifferent
each regiontissues are
are not good.
obvious.
There is good density and 3 contrast between
Brightness anddifferent tissues. different
contrast between Brightness is uniform
tissues are good.across areas, and in-
4
ternal details are visible.4 There is good density and contrast between different tissues. Brightness is uniform
across areas, and internal details are visible.
Panorama images generated from four different methods were evaluated by three
dentists. The rating
Panorama process
images is conducted
generated fromusing blinded method.
four different methodsThe results
were show that
evaluated the
by three
panorama image
dentists. The reconstructed
rating by our proposed
process is conducted methodmethod.
using blinded has the The
highest overall
results showscore
that of
the
11.16 ± 2.64. image
panorama The scores assessed by
reconstructed bythe
ourdentists
proposedaremethod
shown in Figure
has 8.
the highest overall score of
11.16 ± 2.64. The scores assessed by the dentists are shown in Figure 8.

Figure
Figure8.8.Boxplot
Boxplotofofcontrast
contrastscores
scoresfor
forpanoramic
panoramicimages.
images.

3.5.Limitations
3.5. Limitations
Although our proposed method achieves excellent performance on the task of auto-
Although our proposed method achieves excellent performance on the task of auto-
matically reconstructing panoramic images, it also has some limitations. First, only a small
matically reconstructing panoramic images, it also has some limitations. First, only a small
amount of curved MPR images in the dental arch region were extracted for the growth
amount of curved MPR images in the dental arch region were extracted for the growth
shape of the tooth; thus, the internal continuous information collected was insufficient. Ad-
shape of the tooth; thus, the internal continuous information collected was insufficient.
ditional information on the growth shape of the tooth could potentially benefit the display
Additional information on the growth shape of the tooth could potentially benefit the dis-
and further improve the quality of panoramic images. Second, the values of α, β, and γ in
play and further improve the quality of panoramic images. Second, the values of α, β, and
Formulas (3) and (4) may vary due to different models of CBCT scanners, consequently
γ in Formulas (3) and (4) may vary due to different models of CBCT scanners, conse-
increasing the workload. Finally, we have only tested the algorithm on limited data and
quently increasing the workload. Finally, we have only tested the algorithm on limited
have not yet applied it clinically.
data and have not yet applied it clinically.
4. Conclusions
In this paper, we proposed a fast, automatic reconstruction method for panoramic
images based on CBCT. This method first detects the dental arch curve through axial MIP
images over various ranges and then applies an enhanced image fusion algorithm to extract
different tissues and adjust the intensity distribution during the reconstruction process.
Our results indicate that the proposed method effectively reduces the disturbance among
different types of tissue by utilizing sub-regional detection, notably increases the fitting
Electronics 2022, 11, 2404 10 of 11

accuracy of the dental arch curve by implementing a control point adjustment panel, and
significantly brings down the influence of metal implants through intensity distribution
adjustment. Therefore, the dentition display appears on ideal brightness, high level of
completion, and proper contrast in both tooth and treatment areas. Despite the limitations
mentioned in Section 3.5, our work presented a high-speed, reliable, and satisfactory
panoramic image reconstruction technique for digital dentistry.
In future work, we plan to construct new transfer functions for different tissues or
regions of interest during the synthesis of panoramic images or to use different algorithms
to enhance and improve panoramic images [36–38]. In addition, we will try to extract the
dental arch curves at different positions to further improve the performance and generality
of the algorithm.

Author Contributions: Conceptualization, J.Z. and L.S.; methodology, Y.J.; software, Y.J.; validation,
Y.J., J.Z. and L.S.; formal analysis, Y.J.; investigation, F.G.; resources, S.Z.; data curation, F.Y.; writing—
original draft preparation, Y.J.; writing—review and editing, J.Z., L.S. and F.Y.; visualization, Y.J.;
supervision, J.Z.; project administration, J.Z.; funding acquisition, J.Z. All authors have read and
agreed to the published version of the manuscript.
Funding: This research was funded by Natural Science Foundation of Shanghai, grant number
19ZR1455100 and the Natural Science Research Project of Minhang District, Shanghai, grant number
2019MHZ039.
Institutional Review Board Statement: The study was conducted in accordance with the Declaration
of Helsinki, and approved by the Ethics Committee of Shanghai Fifth People’s Hospital, Fudan
University (protocol code: 2019-065, approved: 28 August 2019).
Informed Consent Statement: Informed consent was waived because of the retrospective nature of
the study and the analysis used anonymous clinical data.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Farman, A.G.; Scarfe, W.C. Historical perspectives on CBCT. In Maxillofacial Cone Beam Computed Tomography; Springer: Cham,
Switzerland, 2018; pp. 3–11.
2. Frederiksen, N.L. Diagnostic imaging in dental implantology. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endodontol. 1995, 80,
540–554. [CrossRef]
3. Fokas, G.; Vaughn, V.M.; Scarfe, W.C.; Bornstein, M.M. Accuracy of linear measurements on CBCT images related to presurgical
implant treatment planning: A systematic review. Clin. Oral Implant. Res. 2018, 29, 393–415. [CrossRef] [PubMed]
4. Guerrero, M.E.; Jacobs, R.; Loubele, M.; Schutyser, F.; Suetens, P.; van Steenberghe, D. State-of-the-art on cone beam CT imaging
for preoperative planning of implant placement. Clin. Oral Investig. 2006, 10, 1–7. [CrossRef] [PubMed]
5. Manavella, V.; Romano, F.; Garrone, F.; Terzini, M.; Bignardi, C.; Aimetti, M. A novel image processing technique for 3D
volumetric analysis of severely resorbed alveolar sockets with CBCT. Minerva Stomatol. 2017, 66, 81–90. [CrossRef] [PubMed]
6. Harris, D.; Horner, K.; Gröndahl, K.; Jacobs, R.; Helmrot, E.; Benic, G.I.; Bornstein, M.M.; Dawood, A.; Quirynen, M. EAO
guidelines for the use of diagnostic imaging in implant dentistry 2011. A consensus workshop organized by the European
Association for Osseointegration at the Medical University of Warsaw. Clin. Oral Implant. Res. 2012, 23, 1243–1253. [CrossRef]
[PubMed]
7. Çiftçi, M.E.; Aktan, A.M.; İşman, Ö.; Yıldırım, E. Relationship between CBCT and panoramic images of the morphology and
angulation of the posterior mandibular jaw bone. Surg. Radiol. Anat. 2016, 38, 313–320. [CrossRef] [PubMed]
8. Cui, W.; Zeng, L.; Chong, B.; Zhang, Q. Toothpix: Pixel-Level Tooth Segmentation in Panoramic X-Ray Images based on Generative
Adversarial Networks. In Proceedings of the 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI), Nice,
France, 13–16 April 2021; pp. 1346–1350. [CrossRef]
9. Wei, X.; Wang, Y. Inferior alveolar canal segmentation based on cone-beam computed tomography. Med. Phys. 2021, 48, 7074–7088.
[CrossRef] [PubMed]
10. Warin, K.; Limprasert, W.; Suebnukarn, S.; Inglam, S.; Jantana, S.; Vicharueang, S. Assessment of deep convolutional neural
network models for mandibular fracture detection in panoramic radiographs. Int J Oral Maxillofac Surg 2022, 51. [CrossRef]
[PubMed]
11. Song, W.; Liang, Y.; Yang, J.; Wang, K.; He, L. Oral-3d: Reconstructing the 3d structure of oral cavity from panoramic x-ray. In
Proceedings of the AAAI Conference on Artificial Intelligence, Virtually, 2–9 February 2021; Volume 35, pp. 566–573.
12. Kanitsar, A.; Fleischmann, D.; Wegenkittl, R.; Felkel, P.; Groller, E. CPR-Curved Planar Reformation; IEEE: Manhattan, NY, USA, 2002.
Electronics 2022, 11, 2404 11 of 11

13. Bornstein, M.M.; Horner, K.; Jacobs, R. Use of cone beam computed tomography in implant dentistry: Current concepts,
indications and limitations for clinical practice and research. Periodontol. 2000 2017, 73, 51–72. [CrossRef] [PubMed]
14. Harris, D.; Buser, D.; Dula, K.; Gröndahl, K.; Harris, D.; Jacobs, R.; Lekholm, U.; Nakielny, R.; van Steenberghe, D.; van der Stelt, P.
EAO Guidelines for the use of Diagnostic Imaging in Implant Dentistry: A consensus workshop organized by the European
Association for Osseointegration in Trinity College Dublin. Clin. Oral Implant. Res. 2002, 13, 566–570. [CrossRef] [PubMed]
15. Suzuki, S.; Ichikawa, K.; Tamaki, S. Image Quality and Clinical Usefulness of Ray-summation Image Reconstructed from CT Data,
Compared with Digital Radiography. Nihon Hoshasen Gijutsu Gakkai Zasshi 2017, 73, 372–381. [CrossRef] [PubMed]
16. Westenberg, M.A.; Roerdink, J.B.T.M. X-ray volume rendering by hierarchical wavelet splatting. In Proceedings of the 15th
International Conference on Pattern Recognition, ICPR-2000, Barcelona, Spain, 3–7 September 2000; IEEE: Manhattan, NY, USA,
2000; Volume 3, pp. 159–162.
17. Izzetti, R.; Nisi, M.; Aringhieri, G.; Crocetti, L.; Graziani, F.; Nardi, C. Basic knowledge and new advances in panoramic
radiography imaging techniques: A narrative review on what dentists and radiologists should know. Appl. Sci. 2021, 11, 7858.
[CrossRef]
18. Chanwimaluang, T.; Sotthivirat, S.; Sinthupinyo, W. Automated dental arch detection using computed tomography images. In
Proceedings of the 2008 9th International Conference on Signal Processing, Beijing, China, 6–29 October 2008; IEEE: Manhattan,
NY, USA, 2008; pp. 737–740.
19. Amorim, P.H.; Moraes, T.F.; Silva, J.V.; Pedrini, H.; Ruben, R.B. Reconstruction of panoramic dental images through bézier
function optimization. Front. Bioeng. Biotechnol. 2020, 8, 794. [CrossRef] [PubMed]
20. Zhu, S.; Fang, H.; Zhang, D. An Algorithm for Automatically Extracting Dental Arch curve. J. Phys. Conf. Ser. IOP Publ. 2021,
2082, 012018. [CrossRef]
21. Bae, M.; Park, J.W.; Kim, N. Semi-automatic and robust determination of dental arch form in dental cone-beam CT with B-spline
approximation. Comput. Methods Prog. Biomed. 2019, 172, 95–101. [CrossRef]
22. Akhoondali, H.; Zoroofi, R.A.; Shirani, G. Fully automatic extraction of panoramic dental images from CT-scan volumetric data
of the head. J. Appl. Sci. 2009, 9, 2106–2114. [CrossRef]
23. Bing, H.; Liang, C.; Zhen, C.; Fang, P.; Deyu, L.; Shuyu, L.; Yubo, F. An automatic method of synthesizing panoramic radiograph by
unwrapping dental CT image. In Proceedings of the 2011 International Conference on Mechatronic Science, Electric Engineering
and Computer (MEC), Jilin, China, 19–22 August 2011; IEEE: Manhattan, NY, USA, 2011; pp. 1094–1096.
24. Papakosta, T.K.; Savva, A.D.; Economopoulos, T.L.; Matsopoulos, G.K.; Gröhndal, H.G. An automatic panoramic image
reconstruction scheme from dental computed tomography images. Dentomaxillofacial. Radiol. 2017, 46, 20160225. [CrossRef]
25. Sa-Ing, V.; Wangkaoom, K.; Thongvigitmanee, S.S. Automatic dental arch detection and panoramic image synthesis from CT
images. In Proceedings of the 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology
Society (EMBC), Osaka, Japan, 3–7 July 2013; IEEE: Manhattan, NY, USA, 2013; pp. 6099–6102.
26. Luo, T.; Shi, C.; Zhao, X.; Zhao, Y.; Xu, J. Automatic synthesis of panoramic radiographs from dental cone beam computed
tomography data. PLoS ONE 2016, 11, e0156976. [CrossRef]
27. Yun, Z.; Yang, S.; Huang, E.; Zhao, L.; Yang, W.; Feng, Q. Automatic reconstruction method for high-contrast panoramic image
from dental cone-beam CT data. Comput. Methods Prog. Biomed. 2019, 175, 205–214. [CrossRef]
28. Guo, H. A simple algorithm for fitting a Gaussian function [DSP tips and tricks]. IEEE Signal Process. Mag. 2011, 28, 134–137.
[CrossRef]
29. Hall, R.W. Fast parallel thinning algorithms: Parallel speed and connectivity preservation. Commun. ACM 1989, 32, 124–131.
[CrossRef]
30. Wang, P.S.P.; Zhang, Y.Y. A Fast Serial and Parallel Thinning Algorithm Cybernetics and Systems’ 86; Springer: Dordrecht, The
Netherlands, 1986; pp. 909–915.
31. Zhang, T.Y.; Suen, C.Y. A fast parallel algorithm for thinning digital patterns. Commun. ACM 1984, 27, 236–239. [CrossRef]
32. Amorim, P.H.; de Moraes, T.F.; da Silva, J.V.; Pedrini, H.; Ruben, R.B. Automatic reconstruction of dental CT images using
optimization. In Proceedings of the Biodental Engineering III: Proceedings of the 3rd International Conference on Biodental
Engineering (BIODENTAL 2014), Porto, Portugal, 22–23 June 2014; Taylor & Francis Books Ltd.: Abingdon, UK, 2014; pp. 57–62.
33. Gijbels, F.; De Meyer, A.M.; Bou Serhal, C.; Van den Bossche, C.; Declerck, J.; Persoons, M.; Jacobs, R. The subjective image quality
of direct digital and conventional panoramic radiography. Clin. Oral Investig. 2000, 4, 162–167. [CrossRef] [PubMed]
34. Gamache, C.; English, J.D.; Salas-Lopez, A.M.; Rong, J.; Akyalcin, S. Assessment of image quality in maxillofacial cone-beam
computed tomography imaging[C]Seminars in Orthodontics. WB Saunders 2015, 21, 248–253.
35. Wang, Z.; Bovik, A.C.; Sheikh, H.R.; Simoncelli, E.P. Image quality assessment: From error visibility to structural similarity. IEEE
Trans. Image Process. 2004, 13, 600–612. [CrossRef]
36. Lee, M.S.; Park, C.H.; Kang, M.G. Edge enhancement algorithm for low-dose X-ray fluoroscopic imaging. Comput. Methods Prog.
Biomed. 2017, 152, 45–52. [CrossRef] [PubMed]
37. Zhao, C.; Wang, Z.; Li, H.; Wu, X.; Qiao, S.; Sun, J. A new approach for medical image enhancement based on luminance-level
modulation and gradient modulation. Biomed. Signal Process. Control. 2019, 48, 189–196. [CrossRef]
38. Rahmi-Fajrin, H.; Puspita, S.; Riyadi, S.; Sofiani, E. Dental radiography image enhancement for treatment evaluation through
digital image processing. J. Clin. Exp. Dent. 2018, 10, e629. [CrossRef] [PubMed]