Henry Ford
Health System
RADIOLOGY RESEARCH
Performance for Radiological
Display Devices
Michael Flynn
Dept. of Radiology
mikef@rad.hfh.edu
Projection Test Pattern
12 / 0
243 / 255
243 / 255
12 / 0
iQC Test Pattern (pacsDisplay)
Flynn AAPM 2008
2
Intro: visual interpretation
(A) Subject contrast in the patient is;
(B) recorded by the detector and
(C) transformed to display values that are
(D) and sent to a display device for presentation to
(E) the human visual system and interpretation.
DETECTION
DISPLAY
The device used to display radiographic images must effectively
transfer spatial and contrast information to the human observer.
Flynn AAPM 2008
3
Intro: Visual Requirements
The performance of the human visual system (HVS)
is reviewed in relation to display for the primary
interpretation of radiological images.
A. Viewing Distance
B. Display Size
C. Pixel Size
D. Display Zoom
E. Equivalent Contrast
Flynn AAPM 2008
4
A. Viewing Distance?
•Vergence
•Accomodation
• Vergence (convergence)
allows both eyes to focus
the object at the same
place on the retina.
• The closer the object,
the more the extraocular
muscles converge the
eyes inward towards the
nose.
Flynn AAPM 2008
5
A. Viewing distance and vergence
Resting Point of Vergence



The eyes have a resting point of vergence of about 40
inches.(Jaschcinsk-Kruza 1991).
–
Objects closer than the resting point cause muscle strain.
–
The closer the distance, the greater the strain (Collins 1975).
Every one of the subjects studied by Jaschinski-Kruza
(1998) judged the eye to screen distance of 20 inches
to be too close. All accepted a 40 inch distance.
Grandjean (1983) reported an average preferred
viewing distance of 30 inches.
Arms length viewing distance
Flynn AAPM 2008
6
A. Viewing distance and accomodation

Resting Point of Accommodation
The ciliary muscle changes the shape
of the lens to focus the object.
– The eyes have a resting point of
accommodation which is the
distance that the eye focuses to
when there is nothing to look at
(Owens 1984).
– This resting point averages about
31 inches (Krueger 1984).

Prolonged viewing of a monitor closer than the resting
point of accommodation increases eye strain (JaschinskiKruza 1988). The ciliary muscle must work 2.5 times
harder to focus on a monitor 12 inches away than it does
to focus at 30 inches.
Flynn AAPM 2008
Arms length viewing distance
7
B. Display Size?
Flynn AAPM 2008
Field of view in relation
to viewing distance.
8
B. HVS: peripheral response
Rod
have a
high
The receptors
retina contains
large
sensitivity,
grayreceptors
response,
number of rod
and
that
(160interconnections
M) distributed over
respond to motion.
the peripheral field.
Flynn AAPM 2008
9
B. Display Size vs Viewing Distance
For a specific viewing distance the diagonal dimension
should be about 80% of the viewing distance. (44o)
Task
Viewing Distance
Diagonal Size
Close Inspection
1/3 meter
10.4 inches
Normal viewing
2/3 meter
20.8 inches
Consultation viewing
1 meter
31.5 inches
Teaching Conference
3 meters
110.1 inches
Flynn AAPM 2008
10
B. Field of View
21 inch (diagonal) monitors with a field of 32 x 42 cm
provide an effective field for radiographic images
viewed at a normal distance (2/3 m).
Eyeglass
lens
should be
optimized
for a
normal
viewing
distance
Flynn AAPM 2008
11
C. Pixel Size?
•Visual Acuity
•Contrast Sensitivity
Flynn AAPM 2008
12
C. Visual Acuity
A variety of test patterns are used to assess visual
acuity. Clinical measures are done typically with a
Snellen eye chart. Much psychovisual research has
been done using sinusoidally modulated test targets.
Flynn AAPM 2008
13
C. HVS: Retinal anatomy
The retina of the human eye contains a network of
rods and cones interconnected by neural cells.
Flynn AAPM 2008
14
C. HVS: Foveal response
Particularly
cones
(2 m)
At 60 cm, 1 thin
degree
corresponds
are
packed
in the
to adensely
1 cm field
of view.
This
central
50
microns
of
the
fovea
area is focused on a 288 micron
centralis.
They
provide
region of the
retina,
thehigh
fovea
detail color response.
Flynn AAPM 2008
15
C. Contrast Sensitivity as a measure of spatial acuity
Contrast sensitivity is the inverse of contrast threshold: Cs = 1/Ct
~0.5 c/mm
~2.5 c/mm
10% max
Flynn AAPM 2008
16
C. Spatial Frequency: cycles/degree
• The eye perceives luminance variations as a change
with respect to viewing angle.
mm

• Data on visual performance must always be converted from
cycles/degree to cycles/mm at a specified viewing distance.
Cycles/mm = 57.3 x (cycles/degree) / (viewing distance, mm)
Flynn AAPM 2008
17
C. Pixel Size at Maximum Spatial Acuity


The visual spatial frequency limit and associated pixel size can
be defined as that for which Cs = 10% of maximum.
The pixel size of a display system that matches the resolving
power of the human eye depends on the observation distance.
Flynn AAPM 2008
Distance
frequency
pixel size
Close inspection
(0.33 m)
5 cycles/mm
0.100 mm/pixel
Normal viewing
(0.66 m)
2.5 cycles/mm
0.200 mm/pixel
Consultation view
(1.00 m)
1.7 cycles/mm
0.300 mm/pixel
Conference room
(3.00 m)
0.5 cycles/mm
1.000 mm/pixel
18
C. Pixel array and Megapixels


The pixel size and the field of view dictate the pixel
array size and the total number of pixels.
Megapixels alone is not a good descriptor of quality.
Field of View
pixel size
array size
MegaPixels
21 inch
0.100 mm
3200 x 4200
13.4
21 inch
0.200 mm
1600 x 2100
3.4
• idtech 3 MP panel
20.8 inch (32 x 42 cm) 3.1 megapixels (.207 mm pixels)
Flynn AAPM 2008
19
C. LCD 2MP Colot Pixel


LCD Pixel Structure
For a pixel pitch greater than ~200 microns, the pixel
structure is visible as a granular pattern.
Some consumer monitors have a granular diffusing
surface that creates a random noise pattern.
Dual Domain pixel structure
Flynn AAPM 2008
Single Domain pixel structure
20
D. Display Zoom?
Detector Detail in relation to
Display Acuity
Flynn AAPM 2008
21
D. Viewing distance and image zoom


Use of image zoom features is ergonomically better than
leaning forward for close inspection.
Split deck tables with a broad front deck usefully prohibit
close inspection with 3 MP monitors.
Flynn AAPM 2008
22
D. Magnification / Minification
4X
1/4X
1X
Zoom is needed to display detail at
the detector pixel level with good
contrast sensitivity.
Flynn AAPM 2008
1X
Minification has value by
increasing the frequency of
diffuse structures.
23
D. True Size
For some applications, “true size”
display is important.
–
Comparison of current and
prior exams obtained on
different detectors (or with
screen-film).
–
Orthopedic assessment of
size.
This requires knowledge of
–
Detector element (del) pitch
–
Display element (pixel) pitch.
Flynn AAPM 2008
Prior
Current
* adapted from D. Clunie, SCAR 2005
24
D. Re-sampling for Display
DETECTOR
A subset of image
values is re-sampled
for presentation on a
display device.
DISPLAY
In General;
• The detector and display
pixel spacings are different.
• The detector and display
overall size are different
Flynn AAPM 2008
25
D. Up-sampling (magnification)


Up sampling occurs
when the number of
display values in the
region re-sampled is
more than the number
of recorded image
values .
This is commonly
encountered when
displaying CT and MR
images.
• Blue circles show an 11x11
array of recorded image
pixel values.
• Green solid circles are for a
15 x 15 array of display
pixel values
Flynn AAPM 2008
26
D. down-sampling (minification)


Down sampling occurs
when the number of
display values in the
region re-sampled is
less than the number
of recorded image
values .
This is commonly
encountered when a
full radiograph is
displayed.
• Blue circles show an 11x11
array of recorded image
pixel values.
• Green solid circles show a
7 x 7 array of display values.
Flynn AAPM 2008
27
D. Interpolation
Estimation of variably spaced display
values from a set of image values is done
using mathematical interpolation methods.
Flynn AAPM 2008
28
D. Approximate Interpolation
While fast, nearest neighbor and bi-linear interpolation do not
result in optimal image quality due to artifacts and blur.
Nearest Neighbor Interpolation
Bi-Linear Interpolation
• Display value (green) is taken as the
image value (blue) at the nearest row
and column.
• Image values pairs above & below the
display value are linearly interpolated
based on the column position (black).
• Produces visible block artifacts for
large magnification.
• These values are linearly interpolated
based on the row position.
Flynn AAPM 2008
29
D. Improved Interpolation
Improved quality can be achieved by
estimating display values from the
closest 16 image values (4 x 4).
– Spline interpolation uses polynomial
arc segments constrained to be
smooth (1st and 2nd derivatives) at
transition points (nodes). It has
been classically used for digital
images.
– A still popular technique known as
cubic convolution involves the use of
a sinc-like kernels composed of
piecewise cubic polynomials.
– Recent work has shown that
generalized spline interpolation using
a pre-filter operation provides
excellent performance with fast
implementation and can provide
controlled smoothing.
Flynn AAPM 2008
Cubic Interpolation
• Display value (green) is computed
from the closest 16 image values.
• The weighting functions for the 16
image values are intended to estimate
a continuous function within the space
between the sampled values.
30
D. Magnification / Minification
Magnification: Calcified duct, 4:1 re-sampling 5.25 x 5.25 mm region
A
Nearest Neighbor
B
Bi-Linear
C
Cubic
Minification.
• Advanced interpolation methods can also provide effective
minification with noise reduction (low-pass filter).
• Alternatively, minification is often done using multi-scale
representations of the image with progressive presentation.
Flynn AAPM 2008
31
D – Display Interpolation – key points
Interpolation and Image Quality:


The numerical approach used to obtained magnified
display values has significant impact on image quality.
Modern interpolation with good performance needs
optimal implementation for high speed.
Minification and noise reduction:


Minification should be done such that high frequency
noise (quantum mottle) is reduced.
Multi-scale representation of image date provides a
means for minification (JPEG 2000, JPIP, Wavelet, ..).
Flynn AAPM 2008
32
E. Equivalent Contrast?
• Grayscale response
• Luminance ratio (L’max/L’min)
Flynn AAPM 2008
33
E. Contrast detection in relation to brightness
• Contrast detection is diminished for images with low brightness.
• Extensive experimental models have documented the dependence
of contrast detection on luminance, spatial frequency, orientation
and other factors. The empirical models of either S. Daly or J.
Barton provide useful descriptions of this experimental data.
Flynn AAPM 2008
34
E. Contrast threshold vs luminance
Flynn AAPM 2008
The Barton model describes the average contrast
threshold of normal observers. Significant differences
exist for individual observers for different test methods
35
E. DICOM graylscale display standard
DICOM part 3.14 describes a grayscale response that
compensates for visual deficits at low brightness
Flynn AAPM 2008
36
E. Fixed versus variable adaptation
Contrast threshold for varied visual adaptation (A, Flynn 1999b) and fixed
(B) visual adaptation: The contrast threshold, L/L, for a just noticeable
difference (JND) depends on whether the observer has fixed (B) or varied
(A) adaptation to the light and dark regions of an overall scene.
Flynn AAPM 2008
37
E – Ct for small sinusoidal patterns on a color LCD.
SINE and ADAPT Contrast Thresholds Normalized to SINE
4.5
4
DB
3.5
Relative CT/CBM
2AFC
assessment of
Ct using varied
background
region
brightness.
DP
3
MF
MP
2.5
PT
2
PR-80
SL
1.5
AVERAGE
1
0.5
0.1
1
10
100
L/L_SINE
Flynn AAPM 2008
Effects of adaptation on observers contrast
thresholds relative to changes in background.
38
E. Adapted Observer Performance
Observer performance is best when visual system is
adapted to the average scene luminance.
A
Flynn AAPM 2008
B
C
39
E. Effect of Lmax/Lmin


Digital radiographs
should be displayed
using over a
luminance range of
250-350:1.
Images prepared
for range of 250
that are display on
a monitor with large
range will have
poorly perceived
contrast in dark
regions.
Flynn AAPM 2008
250:1  .1 to 2.50 OD
350:1  .1 to 2.65 OD
650:1  .1 to 2.90 OD
250:1
650:1
40
E. LR for LCD monitors



Flynn AAPM 2008
For CRT monitors, LR is set by adjusting
brightness (Lmin) and contrast (Lmax).
For LCD devices, only the backlight
intensity can be adjusted.
For LCD devices
–
Lmax is set by adjusting the backlight
brightness (current control).
–
Lmin is set as a part of the grayscale calibration
(starting LUT value).
41
Other issues
Issues that I have not addressed!
• LCD devices have significant contrast changes
when viewed at angles oblique to the surface.
• Note: New OLED technologies promise to
eliminate that problem in near future.
• Pixel noise is poorly documented for new LCD
monitors. Further works needs to be done to
understand whether pixel noise effects diagnostic
visual performance.
• 256 (8bit) versus 1024 (10bit) gray levels.
Flynn AAPM 2008
42
Questions?
Flynn AAPM 2008
43