Artifacts In Digital Radiography

Academic Paper Writing Service
http://StudyHub.vip/Artifacts-In-Digital-Radiography 👈

1. ARTIFACTS IN DIGITAL RADIOGRAPHY
DAVIDA. JIME´NEZ,LAURAJ. ARMBRUST,ROBERTT. O?BRIEN,DAVIDS. BILLER
Digital radiography is becoming more prevalent in veterinary medicine, and with its increased use has come the
recognition of a number of artifacts. Artifacts in digital radiography can decrease image quality and mask or
mimic pathologic changes. They can be categorized according to the step during which they are created and
include preexposure, exposure, postexposure, reading, and workstation artifacts. The recognition and under-
standing of artifacts in digital radiography facilitates their reduction and decreases misinterpretation. The
purpose of this review is to name, describe the appearance, identify the cause, and provide methods of resolution
of artifacts in digital radiography.Veterinary Radiology & Ultrasound, Vol. 49, No. 4, 2008, pp 321–332.
Key words: artifact, computed, digital, radiography.
Introduction
W
ITH INCREASING USE of digital radiography in veter-
inary medicine, numerous artifacts are being iden-
tified. Artifacts include any alteration in the image, which
leads to a misrepresentation or hindered visualization of
the object of interest. These artifacts must be recognized to
correctly interpret digital radiographs. An understanding
of how an artifact corresponds to a given step during image
acquisition and display will facilitate implementation of an
expedient remedy. Hardware and software troubleshooting
in digital radiography differs from troubleshooting in film-
screen radiography, and these differences must be under-
stood.
Digital radiography includes all forms of radiography
that incorporate a digitally processed image. These include
photostimulable phosphors, direct digital radiography, and
indirect digital radiography. Photostimulable phosphor
systems are marketed as computed radiography (CR), and
the latter two are marketed as digital radiography (DR).
These systems share a number of steps in common during
the imaging process. However, some unique features of
each imaging system are a source of artifact seen only with
the corresponding system.
Digital radiographic artifacts can be divided into the
following categories corresponding to the step during
which they are created: preexposure artifacts, exposure ar-
tifacts, postexposure artifacts, reading artifacts, and work-
station artifacts. Identifying artifacts using standardized
nomenclature facilitates the continued understanding of
digital radiography as it becomes more prevalent in vet-
erinary medicine. The purpose of this review is to name,
describe the appearance, identify the cause, and provide
methods of resolution of artifacts in digital radiography.
Preexposure Artifacts (Table 1)
Storage Scatter
The highly sensitive imaging plates used with CR sys-
tems are more likely than conventional film to be notice-
ably affected by extraneous radiation exposure.
1–3
Imaging
plate fog due to scatter or background radiation can occur
before or after exposure.
4
Fog is often due to exposure to
scatter radiation during other diagnostic imaging studies
(Fig. 1). Storage scatter leads to an overall decrease in
image quality and contrast. Objects between the source of
extraneous radiation and the imaging plate can obstruct a
portion of the scattered radiation and be present on the
image as an area devoid of added exposure.
5
This may
occur with cassette storage devices or shelving units. An
imaging plate should be kept outside of the imaging suite
during any radiographic acquisition for which it is not
used. Background radiation exposure of the highly sensi-
tive imaging plates is often unavoidable. Erasure of imag-
ing plates before use is recommended to eliminate storage
scatter and is often done at the beginning of each work-
day.
4
Cracks
With CR systems, repeated transport of imaging plates
through the imaging plate reader makes them susceptible
to physical damage. Inadequate maintenance and inap-
propriate handling may also be a cause of imaging plate
damage.
6
Cracks in the imaging plate appear as white lines
or dots that are usually located at the edge of the im-
age.
4,5,7,8
Cracks closer to the center of the imaging plate
are more likely to be superimposed over a region of inter-
est, and may mimic linear mineral opacities.
7
Reducing the
Address correspondence and reprint requests to Dr. David A. Jime´nez,
at the above address. E-mail: [email protected].
Received June 20, 2007; accepted for publication January 21, 2008.
doi: 10.1111/j.1740-8261.2008.00374.x
From the Department of Radiology and Diagnostic Imaging, College of
Veterinary Medicine, Kansas State University, Manhattan, KS 66506.
321

2. phosphor grain size, phosphor layer thickness, and surface
protection layer are preferred for higher image resolution,
but these characteristics make the imaging plate more
likely to be physically damaged.
5
Scheduled maintenance
of the cassette, imaging plate, and imaging plate reader
reduces the production of cracks. Imaging plates must be
handled carefully with cotton gloves. Imaging plates with
cracks that interfere with the image should be replaced.
8
Partial Erasure
Imaging plates retain a portion of the latent image after
being read. They are subsequently erased by light exposure.
CR images created by exposure to high quantities of
X-rays have a more intense latent image, which may persist
after the imaging plate is read. In these instances, a greater
amount of stored energy remains within the imaging plate.
As such, routine exposure to visible light may be insuffi-
cient to completely erase the imaging plate.
5
Technical
malfunction is another cause of incomplete imaging plate
erasure and may be due to electrical malfunction or burnt
out, fading, or dirty light bulbs. Partial erasure is more
likely with newer imaging plates, which have more stable
energy storage.
5
When a partially erased imaging plate is
subsequently used for a radiographic study, the latent im-
age remains on the imaging plate. It appears as a faint
image superimposed over the more recent radiograph.
4
Bright visible light is most commonly used for erasure and
should be inspected for proper function. Ultraviolet light
can be used in addition to visible light for more thorough
erasure.
5
Phantom Image
Shortly after imaging plate erasure, the imaging plate
does not have a detectable latent image. However, if the
imaging plate is not read or erased for several days or
longer, the latent image from the previous radiograph may
become increasingly detectable.
8
The previous image ap-
pears faintly superimposed over the more recent image.
When using a CR system, erasure of imaging plates is rec-
ommended before use and can be performed at the begin-
ning of each workday.
7,8
Exposure Artifacts (Table 2)
Quantum Mottle
DR and CR allow for a wider range of exposures that
can produce a diagnostic image, when compared with film-
screen radiography. However, using an appropriate expo-
sure technique remains important. When too few X-rays
are incident on a detector or imaging plate, a grainy image
may result (Fig. 2).
1
The degradation in image quality is
due to increased noise, predominantly quantum mottle,
relative to the number of X-rays detected.
9
Quantum mot-
tle is more prevalent when there is a low signal-to-noise
Ta b l e1. Preexposure Artifacts
Artifact Hardware Appearance Cause Remedy
Storage scatter CR Decreased overall intensity
Decreased image quality
May show pattern of exposure
Exposure to scatter and
background radiation
Erase imaging plates daily
Protect from scatter radiation
Cracks CR White lines or dots
Often in periphery
Physical damage to imaging plate
Repeated stress or inappropriate
handing
Handle imaging plates carefully
Perform scheduled maintenance and cleaning
Replace imaging plates as needed
Partial Erasure CR Faint superimposition of
previous image
Erasure light failure
Incorrect erasure light intensity
Replace erasure lights as needed
Incorporate additional ultraviolet light
erasure phase
Phantom Image CR Faint superimposition of
previously erased image
Prolonged time between erasure and
subsequent exposure
Erase imaging plates daily
CR, computed radiography.
Fig. 1. Storage scatter. A cassette was present near a patient in prepa-
ration for a horizontal radiograph. During exposure, scatter radiation was
attenuated by the vertical rod of the cassette stand (arrow) and the back of
the cassette before being incident upon the imaging plate. Remedy: Only the
cassette utilized for the current radiograph should be present in the imaging
suite during exposure. 5559 mm (300300 DPI)
322 JIMEŁNEZ ET AL. 2008

3. ratio (S:N) and may be due to the production of too few
X-rays or high attenuation of X-rays by the object
or grid.
10
The S:N increases as the number of X-rays de-
tected increases, because noise increases as a square root
function of signal.
1,11
Quantum mottle can be minimized
by increasing the number of X-rays incident on the detec-
tor, while avoiding overexposure. To minimize quantum
mottle artifact, technique charts should be used with set-
tings specific for tissue thickness and the region of interest.
Manufacturer?s recommendations should be followed to
maintain clinical technique factors (S number, IgM value,
exposure index) within the preferred range. Intentional
image blurring, through the use of filtering, is a method
used to reduce the appearance of quantum mottle at the
expense of image detail. Conversely, increasing image de-
tail through image sharpening techniques increases the
perception of quantum mottle. Image sharpness should be
maintained at a reasonable level.
9
Paradoxical Overexposure Effect
Within the diagnostic X-range, an increase in the num-
ber of incident X-rays results in a darker image. This is not
necessarily true for severely overexposed images, in which
areas of greater X-ray exposure may appear increasingly
lighter. This artifact has been observed using indirect dig-
ital radiography, and its mechanism has not been deter-
mined. Paradoxical overexposure effect often manifests
only in areas of minimal X-ray attenuation, such as the
periphery of a region of interest (Fig. 3). Similar exposure
may be maintained by using technique charts with lower
milliampere-seconds and compensatory increased kilovolt-
age peak. If needed, decreasing exposure settings may re-
duce this artifact, but can result in underexposure of the
region of interest.
Table2. Exposure Artifacts
Artifact Hardware Appearance Cause Remedy
Quantum mottle DR
CR
Grainy image
Poor overall image quality
Underexposed image
Insufficient X-ray exposure Increase exposure technique as needed
Paradoxical
overexposure
effect
DR Overexposed areas appear lighter Severe overexposure Reduce exposure technique if possible
Planking DR Rectangular areas of differing
intensity
Variable amplification in separate
sections
Reduce exposure if possible
Upside-down
cassette
CR Cassette backing superimposed on the
image
Severe overall attenuation if
cassette has lead backing
Incorrect cassette placement for
exposure
Check for correct cassette orientation
prior to exposure
Grid cutoff DR
CR
Regions of increased attenuation Incorrect grid placement/
alignment
Check for correct grid location,
orientation, and distance from
anode
Backscatter CR Fogging and decreased image quality
in area of underlying scattering
structure
Scatter from object below the
cassette
High exposure setting
Gap between scattering object
and cassette
Use cassettes with lead backing
Double exposure CR
DR
Images from two exposures
superimposed on each other
Taking multiple radiographs
without erasing the cassette
Memory or transfer errors
Read imaging plates after every
exposure
Provide reliable mode of
data transfer
Dead pixels DR White dots or lines Nonfunctional detector element(s) Use smoothing algorithms for few
affected pixels
Replace detector as needed
CR, computed radiography; DR, digital radiography.
Fig. 2. Quantum mottle. Dorsopalmar radiographs of the right manus of
a dog. (A) The entire image has a grainy appearance resulting in an overall
decrease in image quality. Remedy: (B) The radiograph was repeated after
increased exposure settings. The grainy appearance is dramatically reduced,
and image quality is improved. Note, the area surrounding the manus is
more exposed.
323DIGITALART IFAC TSVol.49,No.4

4. Planking
DR systems can use multiple amplifiers for individual
sections of the digital array. In overexposed regions of the
image, sections using different amplifiers may appear as
sharply demarcated rectangles of differing shades of gray
(Fig. 3). This is referred to as planking and occurs due to
variations in signal amplification that are exacerbated by
high levels of exposure. Recalibration of the detector array
can be performed to more uniformly amplify the X-ray
signal. Using processing techniques that automatically de-
crease exposure values exceeding a predetermined level
may decrease the appearance of artifacts associated with
overexposure.
Upside-Down Cassette
The CR imaging plate can be exposed through either of
its surfaces. However, the cassette must be placed with a
specific surface facing the X-ray source, which is referred to
as the front of the cassette. If the cassette is placed upside-
down during exposure, X-rays will pass through the back
of the cassette before being incident upon the imaging plate.
Attenuation of X-rays will be represented by a well-defined
pattern corresponding to the construction of the back of
the cassette (Fig. 4).
4,7
The artifact created by upside-down
placement of the cassette is similar to that seen with film-
screen radiographs. Correct orientation of the cassette be-
fore each exposure will avoid upside-down cassette artifact.
Grid Cutoff
The increased sensitivity of the imaging plate or detector
makes the use of a scatter reduction grid recommended for
most computed and digital radiographs.
7
Incorrect orien-
tation or position of the grid relative to the X-ray source
will cause attenuation of primary X-rays. An incorrectly
positioned focused grid will create an attenuation pattern
on the image similar to that described in film-screen radi-
ography (Fig. 5).
11
The wider dynamic range in DR and
CR may allow for uniform alterations in X-ray attenuation
without noticeable image degradation. Decentering of a
nonfocused grid may not be apparent on digital or com-
puted radiographs, but patterns of nonuniform attenuation
remain readily recognizable. The ability to easily magnify
digitized images may also make grid lines more apparent.
Correct placement of scatter reduction grids will reduce
these artifacts.
Fig. 3. Paradoxical overexposure effect and planking. Dorsopalmar ra-
diograph of the right distal phalanges of a rhinoceros. The more exposed
background is lighter than the hoof wall (asterisk) and the adjacent needle
(arrow) used to locate a fistulous tract. Well-demarcated rectangles (arrow-
heads) of varied shades of gray are present within the background. Remedy:
Reduce exposure technique if possible.
Fig. 4. Upside-down cassette. Right lateral radiograph of the thorax of a
dog. The cassette used for this radiograph has a backing made of plastic. Its
structure is superimposed over the image as a series of circles (arrows) and
lines (arrowheads) of increased attenuation. Remedy: Orient cassette cor-
rectly before each exposure.
Fig. 5. Grid cutoff. Mediolateral radiographs of a dog crus using a fo-
cused grid. (A) With the grid oriented upside-down, two borders of the
image show increased attenuation, which is more severe toward the periph-
ery (arrows). Remedy: (B) Removal and correct replacement of the grid
eliminates grid cutoff and allows for more uniform exposure.
324 JIMEŁNEZ ET AL. 2008

5. Backscatter
High-imaging plate sensitivity makes CR more suscep-
tible to the effects of exposure due to backscatter, as com-
pared with film-screen radiography.
1,2,5,7
A fraction of the
X-rays that transmit through the imaging plate strike a
distant object. This is a source of scatter radiation that can
return toward and expose the imaging plate. The resultant
exposure may be in the shape of the object causing back-
scatter. Partial attenuation of the backscatter by the cas-
sette may also create a corresponding pattern. The image is
fogged except for those areas in which the back of the
cassette attenuates scatter. Backscatter is most noticeable
when a gap is present between the imaging plate and the
source of backscatter.
2
Backscatter increases with increased
object thickness, field of view, and kilovoltage peak.
1,2
Avoiding excessive exposure techniques and appropriately
collimating the field of view reduces backscatter.
1
Back-
scatter artifacts can be eliminated by using cassettes with
lead backing. This decreases transmission of primary
X-rays and blocks backscatter.
1,2,4,5,7
Double Exposure
Double exposures can occur due to repeated exposure of
a CR imaging plate before it being read or erased.
1,5,10
It
can also occur with most DR systems due to power inter-
ruptions or communication errors.
10
The wide dynamic
range of CR and DR allows for satisfactory display of
images that are exposed to large amounts of X-rays. Dou-
ble exposures often do not appear overexposed. The prom-
inence of one superimposed image compared with the other
is relative to the quantity of incident X-rays during each
image acquisition (Fig. 6).
1,5
Double exposures can be
avoided by reading and erasing imaging plates after each
exposure, providing a reliable power supply, and using a
secure mode of data transfer.
Dead Pixels
In DR, detector arrays have multiple detector elements
that are recognized as nonfunctional when manufactured.
They are referred to as dead pixels.
11
Blurring techniques
are programmed to compensate for only the faulty detector
elements. With time and use, additional detector elements
may fail. The dead pixels may appear as white spots su-
perimposed over the same location on every image. When
an excessive number of dead pixels are present, the detector
array can be recalibrated so as to blur the nonfunctional
detector elements. If image quality becomes unacceptably
decreased due to dead pixels and the blurring techniques
subsequently employed, the detector array may require re-
placement. Detector element failure may also lead to loss
of an entire column of data depending on the method of
detector array readout and nature of the malfunction.
Correction for loss of a column of data due to dead pixels
requires different interpolation techniques or detector ar-
ray replacement.
11
Postexposure Artifacts (Table 3)
Fading
X-rays incident upon a CR plate excite the photostimul-
able phosphor within it, which is usually europium-doped
barium fluorohalide.
11
The amount of molecule excitation
is proportionate to the number of X-rays striking the plate,
and the distribution of excited molecules defines the latent
image.
11
After prolonged periods of time, the excited mol-
ecules return to a neutral state. Fading occurs due to the
gradual decrease in molecule excitation, resulting in a
lighter image of decreased quality (Fig. 7).
5
Noticeable
fading can be produced with storage of exposed imaging
plates for several days before image reading. Imaging
Fig. 6. Double exposure. Right lateral thoracic radiograph of a dog. Two
separate exposures of the thorax were made using appropriate exposure
techniques and are superimposed over each other. The wide dynamic range
of digital radiography avoids the double exposure from appearing overex-
posed. Remedy: Read imaging plates after each exposure.
Table3. Postexposure Artifacts
Artifact Hardware Appearance Cause Remedy
Fading CR More white radiograph
Decreased overall image quality
Prolonged time between
exposure and reading
Read imaging plates shortly after
exposure
Light leak CR Partial image erasure
Decreased image quality
Imaging plate subjected
to ambient light
Read imaging plates shortly after exposure
Avoid handling imaging plates before reading
Exercise proper cassette maintenance
CR, computed radiography.
325DIGITALART IFAC TSVol.49,No.4

6. plates with more stable phosphors are less susceptible to
lossesinstoredenergyovertime.
5
All imaging plates
should be processed shortly after exposure to decrease the
possibility of fading.
Light Leak
After the reading process, CR systems erase imaging
plates by exposure to bright light. If the imaging plate is
subjected to ambient light between X-ray exposure and
reading, it can be erased. The degree of erasure is depen-
dent upon lighting conditions and the duration of light
exposure. As noted with imaging plates subjected to vary-
ing light intensities for several minutes, there is a decrease
in S:N and overall image quality.
3
The portion of the
imaging plate subjected to light leak appears lighter due to
partial release of stored energy in these areas (Fig. 8). More
stable imaging plates are less susceptible to this artifact.
The effect is opposite of that expected for film-screen ra-
diography, in which light exposure increases optical den-
sity. A light leak artifact may be created when addressing a
jammed, unread imaging plate in an imaging plate reader.
Opening a cassette before reading or a damaged cassette
that does not close properly can be a source of light leak.
Light leak artifacts can be minimized by promptly reading
imaging plates after exposure. Routine cassette and imag-
ing plate reader maintenance is recommended.
Reading Artifacts (Table 4)
Debris
The appearance of debris associated artifacts using film-
screen and CR are similar.
1,8
An important difference lies
in the process of their creation. Imaging plates are exposed
directly by X-rays, without the use of intensifying screens.
Debris within the cassette during exposure usually does not
noticeably attenuate X-rays. However, if debris is present
on the imaging plate during the reading process, it can
block light emitted from the imaging plate from reaching
the photomultiplier tube.
7,8
The area over which light is
blocked is processed as being devoid of stored energy and
is displayed on the radiograph as a white area in the shape
of the debris (Fig. 9).
4,7,8
Routine imaging plate cleaning
and removal of identified debris will reduce these artifacts.
The imaging plate manufacturer should be consulted for
cleaning guidelines. Cleaning schedules often depend on
the nature of imaging plate use and storage.
4,7
Dirty Light Guide
In CR, light emitted from the imaging plate during the
reading process can also be blocked within the light guide.
Fig. 7. Fading. Mediolateral radiographs of the shoulder of a dog. Two
radiographs were taken using the same exposure technique. (A) The first
cassette was stored in an area protected from extraneous radiation for 1 week
before being read. There is nonuniform information loss throughout the
image. The image is grainy and there is an overall decrease in quality. Rem-
edy: (B) Radiographs read immediately after exposure do not exhibit fading.
Fig. 8. Light leak. Dorsoventral radiographs of the cervical spine of a dog
during myelography. (A) Before the imaging plate was read, half was ex-
posed to ambient light for approximately four minutes (asterisk). Its partial
erasure is evidenced by overall lightening, decreased image quality, and a
grainy appearance. The opposite half of the radiograph was unaffected.
Remedy: (B) Reading the imaging plate shortly after each radiographic study
and avoiding exposure to ambient light prevents light leak artifact.
326 JIMEŁNEZ ET AL. 2008

7. Dirt on the light guide can block emitted light from reach-
ing the photomultiplier tube in a line along the length of
the image.
4,7
It appears as a white line oriented in the
direction of imaging plate movement during the reading
process (Fig. 10).
7,10
It is usually recommended that the
light guide be cleaned monthly. Imaging plate reader
cleaning and maintenance should be performed routinely
and as needed to reduce this artifact.
4
Skipped Scan Lines
The imaging plate is struck by a laser and read one line
at a time. The placement of each line in a CR image is
reliant upon the continuous movement of the imaging plate
through the reader at a constant, predetermined speed.
Any interruption in the continuous motion of the imaging
plate can cause scan lines to be skipped or distorted.
8
This
artifact results in image foreshortening due to omission of a
short segment of information and has also been referred to
as zipper artifact.
1
Skipped lines are perpendicular to the
direction of image plate motion. This can be due to phys-
ical jarring of the image plate reader or an abrupt alter-
ation in supplied power.
1
The width of omitted image lines
may be represented by a white band parallel to the skipped
scan lines at the trailing border of the image.
1
Providing
a stable, reliable power supply and avoiding jarring the
imaging plate reader will decrease skipped scan lines.
Unequal Phosphors
During the reading process, an exposed imaging plate
releases stored energy in the form of light to create a CR
Fig. 9. Debris. Mediolateral radiograph of an antebrachium (A) and do-
rsopalmar radiograph of a manus (B) of a dog. A small piece of debris was
adhered to the imaging plate during reading of both of the radiographs. A
small white area (arrows) was present on the radiographs until the imaging
plate was cleaned. Remedy: Practice routine image plate maintenance and
clean as needed.
Fig. 10. Dirty light guide. Craniocaudal radiograph of the left stifle of a
dog. Dirt on the light guide creates a white line along the length of the image
(arrows) in the direction of image plate movement. Remedy: Practice routine
image plate reader maintenance and clean the light guide as needed.
Table4. Reading Artifacts
Artifact Hardware Appearance Cause Remedy
Debris CR White dots or lines Debris blocks emitted light from
imaging plate during reading
Clean imaging plates routinely and as
needed
Dirty light guide CR White line in the direction of imaging
plate movement during reading
Dirt on light guide blocks laser from
striking the imaging plate
Perform scheduled reader
maintenance
Clean light guide as needed
Skipped scan lines CR Omission of image information
perpendicular to the direction of
image plate movement during
reading
Abrupt movement of imaging plate
during reading
Power fluctuation
Avoid contact with reader during
imaging plate reading
Provide reliable power supply
Unequal
phosphors
CR Lightened image
Decreased overall image quality
Use of imaging plates and image plate
reader of differing peak wavelength
Use imaging plates and readers
designed to be used together
CR, computed radiography.
327DIGITALART IFAC TSVol.49,No.4

8. image. The wavelength of this light has a peak intensity
characteristic of the energy storage molecules in the imaging
plate. The imaging plate reader will function most effi-
ciently when the reading laser wavelength is equal to the
wavelength of imaging plate peak intensity.
9
Mismatched
platesandreadersmayleadtooverallimagelighteningor
degradation.
5
Even though unequal phosphor artifact has
not been produced experimentally, it remains recommended
that imaging plates of specific peak wavelength are used
only with their corresponding imaging plate readers.
5
Workstation Artifacts (Table 5)
Diagnostic Specifier
Different anatomic regions are best displayed when im-
age-processing parameters are designed for the specific re-
gion of interest.
10
The processing parameters are defined
during CR or DR equipment set-up and can be unique for
each type of study performed. Common parameters in-
clude contrast, brightness, sharpness, edge enhancement,
and density threshold. Incorrect diagnostic specifier selec-
tion may result in suboptimal image processing and radio-
graphic quality (Fig. 11).
5,8
The diagnostic specifier can
digitally label the image respective to the anatomic region,
making incorrect selection readily identifiable.
5,8
Images
can be corrected by reprocessing the image data or by
compensatory manual processing.
5
The correct diagnostic
specifier should be selected before each exposure for de-
sired processing and display.
Moire´Pattern
CR and DR images are read at a sampling frequency
defined by the linear arrangement of pixels by the equip-
Ta b l e5. Workstation Artifacts
Artifact Hardware Appearance Cause Remedy
Diagnostic
specifier
DR
CR
Decreased overall image quality
Poor image contrast
Incorrect region of interest designated
for postacquisition processing
Select correct region of interest for
each exposure
Moire´ DR
CR
Parallel, curved, or wavy lines of
increased attenuation
Interference of grid lines and
sampling frequency
Use oscillating grid
Use grid with high ratio
Align grid perpendicular to sampling
direction
Increase exposure time if needed
Border
detection
DR
CR
Image borders placed within the
region or interest
Incorrect image post-processing
Exposure off-centered on cassette
Highly attenuating objects in area of
interest
Automatic image analysis
Center the area of interest
Use semiautomatic image analysis
Faulty transfer DR
CR
Loss or distortion of the image Loose cable connection
Power fluctuation
Use a reliable transfer method and
power supply
Misplacement DR
CR
Incorrect localization of sections
of the image in the radiograph
Loose cable connection
Power fluctuation
Use a reliable transfer method and
power supply
U¨berschwinger DR
CR
Dark zone surrounding highly
attenuating objects
Unsharp masking techniques used to
accentuate object borders
Use moderate settings and kernel size
to increase object contrast
Density
threshold
DR
CR
Darkening and decreased contrast
of lesser attenuating structures
Inclusion of high-density objects in
histogram analysis and image grayscale
Create density threshold to exclude
metallic objects from histogram analysis
DR, digital radiography; CR, computed radiography.
Fig. 11. Diagnostic specifier. Right lateral radiographs of the abdomen
of a cat. Radiographs were acquired using the same exposure technique. (A)
One radiograph was labeled and processed under the setting for a thorax.
There is poor contrast between tissue of fat and fluid opacity. Remedy: (B)
When labeled and processed under the settings for an abdomen, the image is
correctly displayed, maximizing the radiographic differences between tissues
in the region of interest.
328 JIMEŁNEZ ET AL. 2008

9. ment used. Within the image, a static grid appears as
parallel lines of increased attenuation corresponding to the
grid?s line density. When the grid is angled, the sampling
frequency intersects the grid lines in a regular distribution.
This is more common with grids of low line density that are
mildly divergent from the sampling direction or plate
reader scan lines. The intersection of grid and sampling
frequency creates numerous sequential points of higher
attenuation that are regularly spaced throughout the im-
age.
12
The arrangement of these points may appear as
straight or wavy lines of increased attenuation superim-
posed over the image (Fig. 12).
8,10,12
The lines may be at an
angle different from that of the grid and of differing thick-
ness. This artifact is also referred to as corduroy artifact.
10
It can also be created when imaging any object of regularly
repeated attenuation, such as metallic coils in anesthetic or
monitoring equipment.
8
Moire´artifact can be reduced or
eliminated by the use of an oscillating Potter-Bucky grid.
4,7
However, this artifact has been produced with oscillating
grids when using a short exposure time. Stationary grids
should be oriented perpendicular to the plate reader or
detector array scan lines and be in accordance with the
manufacturer?s specifications.
4,7
Image blurring and band-
stop Gaussian filters have also been employed to reduce
Moire´artifact.
12
However, blurring techniques and filters
usually decrease image detail.
Border Detection
The initial step of processing any DR or CR image is
determining the area of the image to be analyzed. Border
detection can be performed in one of three modes: auto-
matic, semiautomatic, and fixed. Workstations in auto-
matic mode identify collimated borders of the image using
applications such as Agfa?s CR QS Black Border and
ROIfinder software. Image borders may be incorrectly
recognized if the imaging plate and the field of view are
nonparallel by more than 31.
1,5
Other causes of border
detection failure include division of an imaging plate for
multiple exposures, an off-centered object of interest, and
highly attenuating linear objects such as bone or metallic
implants.
10
The image may be incorrectly cropped, placing
a border through the region of interest. Border detection
failure may also result in incorrect image analysis and
cause a decrease in image quality (Fig. 13).
1,5,8
In semiau-
tomatic mode, only a central portion of the image is
included during analysis. The object of interest must be
within the central location for its inclusion in appropriate
image processing. In fixed mode, the computer neither an-
alyzes nor alters the image. Fixed mode can be applied
when variable postexposure processing is not necessary.
Properly centering the object of interest and semiautomatic
image processing is recommended for most studies and re-
duces the frequency of border detection artifact.
Faulty Transfer and Misplacement
Digital and computed radiographs that are acquired
correctly may still appear distorted if there are errors in
image data transfer to the workstation. Multiple pixels or
lines may be altered, missing, or replaced by electronic
noise. The image or a portion of the image can appear
Fig. 12. Moire´. Postsurgical craniocaudal radiograph of the right femur
of a dog. Interference of the grid and sampling frequencies has created lines
of increased attenuation superimposed over the entire image. The thickness
and orientation of the lines does not correspond to that of the grid. Remedy:
Use oscillating grids or stationary grids with high grid ratios oriented per-
pendicular to the scan lines.
Fig. 13. Border detection. Ventrodorsal radiographs of the thoracolum-
bar spine of a dog during myelography. (A) Automatic border detection
methods incorrectly identified the linear increase in attenuation of the spine
as the margin of the radiograph (arrows). The image was cropped to include
the patient?s right half, which was used for image analysis. Remedy: (B)
Reprocessing the image or repeating the exposure with improved alignment
and semiautomatic border detection methods will reduce this artifact.
329DIGITALART IFAC TSVol.49,No.4

10. elongated or can be replaced by bars of varied shades of
gray (Fig. 14). Overall image quality and contrast can also
be decreased. This can result from memory, digitization, or
communication errors.
10
During misplacement, the radio-
graph is fragmented and separate fragments are incorrectly
located within the image (Fig. 15). Fragments correctly
represent the object of interest, but are misplaced in either
a recognizable or seemingly random pattern. The frag-
ments may be multiplied and are often superimposed over
each other. Faulty transfer and misplacement artifacts are
not unique to either hardware or software. Poorly fitted
data transfer cables and fluctuations in power can lead to
intermittently distorted images. Often, immediate repeat
exposure will result in an image without artifact. Appro-
priate workstation set-up, reliable data transfer, and pro-
vision of a steady power supply reduce faulty transfer and
misplacement artifacts. Scheduled preventative mainte-
nance is recommended, and necessary repairs should be
performed if this artifact persists beyond isolated inci-
dences.
U¨berschwinger
The increased dynamic range of DR and CR is often
accompanied by a decrease in image contrast.
10
Image
manipulation can be used to accentuate the margins of
indistinct opacities and improve contrast. This is com-
monly performed by unsharp masking techniques. Un-
sharp masking is an edge enhancement algorithm that
creates a border at the interface between structures of
differing attenuation.
1,7,8,13
The dark border that sur-
rounds highly attenuating objects is referred to as u¨be-
rschwinger, halo artifact, or rebound effect and this has
been recognized in veterinary imaging.
8,13,14
The promi-
nence of the surrounding dark border varies with kernel
size selected during unsharp masking and the difference in
attenuation between adjacent objects.
7
U¨berschwinger is
often present surrounding metallic orthopedic implants
(Fig. 16). It can be differentiated from pathologic changes
because u¨berschwinger is uniform in thickness, conforms
to the shape of the object, and is present at other similar
interfaces. U¨berschwinger becomes less pronounced with
decreases in edge enhancement.
1,13
Minimal use of edge
enhancement is recommended when imaging body regions
containing metallic implants.
Density Threshold
Highly attenuating objects within the image are desig-
nated extremely high pixel values on the DR or CR his-
togram.
7
The image is assigned a long gray scale when
these extreme values are incorporated into image analysis.
The highly attenuating object is assigned the maximum
gray-scale value. The remaining objects in the image are
near the minimum of the gray scale and do not contrast
Fig. 14. Faulty transfer. Right lateral radiograph of the abdomen of a
cat. Longitudinal streaks of alternating shades of gray (arrow) are super-
imposed over the caudoventral aspect of the patient. This was caused by an
error during image transfer due to a loose data cable. Remedy: Subsequent
exposures may be devoid of artifact. Provide a reliable method of data
transfer and power supply.
Fig. 15. Misplacement. (A) Ventrodorsal radiograph of the abdomen of a
dog and (B) ventrodorsal radiograph of the thorax of a cat. Sections of the
images have been incorrectly located within each image. Some sections have
been duplicated and/or superimposed over other sections. Remedy: Subse-
quent exposures may be devoid of artifact. Provide a reliable method of data
transfer and power supply.
330 JIMEŁNEZ ET AL. 2008

11. each other well (Fig. 17). This is referred to as a density
threshold artifact and may be encountered when metallic
objects are present within the image.
5,8
When processing an
image, a maximum and/or minimum pixel value included
in image analysis can be designated.
5,8
Setting a maximum
density threshold to exclude metallic objects will exclude
their corresponding extremely high pixel values from image
histogram analysis. The image will be processed with the
limited inclusion of biologic opacities, optimizing gray
scale and contrast between tissues. The density threshold
should be selected during equipment set-up.
Conclusion
Artifacts may be commonly encountered when using any
of the currently available digital radiography systems. Cas-
sette-based photostimulable phosphor systems implement a
greater number of steps in the imaging process that may be
susceptible to the creation of artifacts. Preexposure, post-
exposure, and reading artifacts are unique to the systems
marketed as CR. Most digital radiography artifacts differ
from those seen with film-screen radiography. Even when
similar, the high sensitivity and increased dynamic range
of digital radiography systems makes them more vulner-
able to certain artifacts and may alter their presentation.
The appearance of many artifacts is attributable to the
preferences chosen during workstation set-up. These pref-
erences should be carefully selected for each type of radio-
graphic study at each workstation. Processing parameters
do not directly alter the image data but do change how the
image data are displayed. Many workstation artifacts can
be corrected without repeating the radiographic study. Ini-
tial and scheduled calibration of digital radiography equip-
ment as well as appropriate scheduled maintenance will
help provide consistent, high-quality images.
8,10
Provision
of a stable, reliable power supply and method of data
transfer is necessary. Faulty software or hardware should
be fixed or replaced.
ACKNOWLEDGMENT
The authors thank Amy C. Dixon-Jime´nez, DVM, for her support
and contribution in reviewing this manuscript.
REFERENCES
1. Solomon SL, Jost RG, Glazer HS, Sagel SS, Anderson DJ, Molina
PL. Artifacts in computed radiography. Am J Roentgenol 1991;157:181–185.
2. Tucker DM, Souto M, Barnes GT. Scatter in computed radiogra-
phy. Radiology 1993;188:271–274.
Fig.16. U¨berschwinger. Craniocaudal radiograph of the stifle of a dog. A
dark halo is present surrounding the metallic implants (arrows). It is of equal
thickness throughout the image and parallels the contour of the implants.
Remedy: Use moderate levels of unsharp masking or process images without
edge enhancement filters.
Fig. 17. Density threshold. Radiographs of a bag of intravenous fluid
were made using 60 mAs and 6.0 kVp before and after inclusion of a lead
glove (asterisks). (A) The unprocessed negative image and (B) processed
positive image of the fluid bag alone demonstrate shades of gray attributed
to the object of interest and appropriate contrast throughout the image. (C)
In the unprocessed negative image of the bag of fluid and lead glove, the fluid
bag is similar in appearance. (D) In the positive image, the lead glove was
included in analysis and image processing. The fluid bag is depicted by very
dark, indiscernible shades of gray and is not identifiable. Remedy: Pro-
gramming a density threshold to exclude high-density, nonbiologic objects
will allow for maximal tissue contrast and image quality.
331DIGITALART IFAC TSVol.49,No.4

12. 3. Ramamurthy R, Canning CF, Scheetz JP, Farman AG. Impact of
ambient lighting intensity and duration on the signal-to-noise ratio of images
from photostimulable phosphor plates processed using DenOptix
s
and
ScanX
s
systems. Dentomaxillofac Radiol 2004;33:307–311.
4. Stearns ED. Computed radiography in perspective. NAVTA J 2004;
Summer: 53–58.
5. Volpe JP, Storto ML, Andriole KP, Gamsu G. Artifacts in chest
radiographs with a third-generation computed radiography system. Am J
Roentgenol 1996;166:653–657.
6. Hammerstrom K, Aldrich A, Alves L, Ho A. Recognition and pre-
vention of computed radiography image artifacts. J Digital Imag
2006;19:226–239.
7. Cesar LJ, Schueler BA, Zink FE, Daly TR, Taubel JP, Jorgenson
LL. Artefacts found in computed radiography. Br J Radiol 2001;74:195–202.
8. Oestmann JW, Prokop M, Schaefer CM, Galanski M. Hardware
and software artifacts in storage phosphor radiography. Radiographics
1991;11:795–805.
9. Huda W, Slone RM, Belden CJ, Wiliams JL, Cumming WA, Palmer
CK. Mottle on computed radiographs of the chest in pediatric patients.
Radiology 1996;199:249–252.
10. Willis CE, Thompson SK, Shepard SJ. Artifacts and misadventures
in digital radiography. Appl Radiol 2004;33:11–20.
11. Bushberg JT, Seibert JA, Leidholdt EM Jr., Boone JM. The Essen-
tial Physics of Medical Imaging, 2nd ed. Philadelphia: Lippincott Williams &
Wilkins, 2002:145–173,255–316.
12. Lin C, Lee W, et al. A study of grid artifacts formation and
elimination in computed radiographic images. J Digital Imaging 2006;
0:1–11.
13. Tan TH, Boothroyd AE. Correspondence: Uberschwinger artefact
in computed radiographs. Br J Radiol 1997;70:431.
14. Hanson SP, Peck JN, Berry CR, Graham J, Stevens G. Radio-
graphic evaluation of the Zurich cementless total hip acetabular component.
Vet Surg 2006;35:550–558.
332 JIMEŁNEZ ET AL. 2008