Conference PaperPDF Available

Abstract and Figures

The presence of depth cues in a visualization can be a great aid in understanding the structure and topology of a vessel tree. Pseudo Chromadepth is a well-known technique for enhancing depth perception in vascular 3D models. Since it strongly relies on the color channel to convey its depth cues, it is traditionally not suited for combined visualizations comprising color-encoded surface parameters. In this paper, we present and evaluate the use of a modified form of Pseudo Chromadepth that supports displaying additional surface parameters using the color channel while still increasing depth perception. This technique has been designed for the visualization of cerebral aneurysm models. We have combined a discretized color scale to visualize the surface parameter with the Pseudo Chromadepth color scale to convey depth using a Fresnel-inspired blending mask. To evaluate our approach, we have conducted two consecutive studies. The first was performed with 104 participants from the general public and the second with eleven experts in the fields of medical engineering and flow simulation. These studies show that Pseudo Chromadepth can be used in conjunction with color-encoded surface attributes to support depth perception as long as the color scale is chosen appropriately.
Content may be subject to copyright.
Combining Pseudo Chroma Depth Enhancement and Parameter
Mapping for Vascular Surface Models
Benjamin Behrendt, Philipp Berg, Bernhard Preim, Sylvia Saalfeld
The definite version of this article is available at http://diglib.eg.org/.
To cite this version:
Benjamin Behrendt, Philipp Berg, Bernhard Preim, Sylvia Saalfeld
Combining Pseudo Chroma Depth Enhancement and Parameter Mapping for Vascular
Surface Models
Proc. of Eurographics Workshop on Visual Computing for Biology and Medicine (EG
VCBM), in print, 2017
Eurographics Workshop on Visual Computing for Biology and Medicine (2017)
S. Bruckner, A. Hennemuth, and B. Kainz (Editors)
Combining Pseudo Chroma Depth Enhancement and Parameter
Mapping for Vascular Surface Models
B. Behrendt1& P. Berg2& B. Preim1& S. Saalfeld1
1Dept. of Simulation and Graphics, University of Magdeburg, Germany
2Dept. of Fluid Dynamics and Technical Flows, University of Magdeburg, Germany
Abstract
The presence of depth cues in a visualization can be a great aid in understanding the structure and topology of a vessel tree.
Pseudo Chromadepth is a well-known technique for enhancing depth perception in vascular 3D models. Since it strongly relies
on the color channel to convey its depth cues, it is traditionally not suited for combined visualizations comprising color-encoded
surface parameters.
In this paper, we present and evaluate the use of a modified form of Pseudo Chromadepth that supports displaying additional
surface parameters using the color channel while still increasing depth perception. This technique has been designed for the
visualization of cerebral aneurysm models. We have combined a discretized color scale to visualize the surface parameter with
the Pseudo Chromadepth color scale to convey depth using a Fresnel-inspired blending mask.
To evaluate our approach, we have conducted two consecutive studies. The first was performed with 104 participants from the
general public and the second with eleven experts in the fields of medical engineering and flow simulation. These studies show
that Pseudo Chromadepth can be used in conjunction with color-encoded surface attributes to support depth perception as long
as the color scale is chosen appropriately.
Categories and Subject Descriptors (according to ACM CCS): I.3.7 [Computer Graphics]: Three-Dimensional Graphics and
Realism—Color, shading, shadowing, and texture I.4.8 [Image Processing and Computer Vision]: Scene Analysis—Depth cues
1. Introduction
A large set of vessel visualization techniques have been developed,
including surface and volume rendering, illustrative techniques and
model-based techniques. Some of these techniques are carefully
adapted to shape or depth perception by using special color scales
[PBC16]. The downside of relying on color scales to convey depth
is that they often prevent the color channel from being used to vi-
sualize parameters on the vessel wall.
Studying vascular diseases, such as plaques and abdominal or
cerebral aneurysms, involves the evaluation of both morphology
and hemodynamic parameters. Morphologic parameters of objects
can usually be directly inferred from the visualization (such as the
position in the vessel tree) or determined by using measuring tools
(such as size or aspect ratio). Hemodynamic parameters are either
measured or derived with a computational fluid dynamic simulation
and either conveyed in the form of stream or path lines (to display
blood flow patterns) or by mapping information directly onto the
vessel surface using a color scale (such as pressure or wall shear
stress). Using colors to encode the information onto the vessel sur-
face is a common technique and well accepted by physicians.
In this paper, we will examine if this type of encoding can
be combined with additional visualization techniques to increase
depth perception. We present a framework for the improved visual-
ization of vessels that features enhanced depth perception in addi-
tion to allowing surface parameters to be mapped to the vessel wall
using color scales. Our approach uses different color scales on the
vessel surface to create separate visualizations for depth and sur-
face parameters, which are then combined using a blending mask.
The generation of this mask is inspired by the Fresnel effect, which
describes the reflection of a surface based on the viewing angle.
An important application for such a technique is supporting the
understanding of cerebral aneurysms. Cerebral aneurysms show a
high prevalence in the western population (3-5 %) [BSB06], while
their annual risk of rupture is below 1 % [MKH12]. On the one
hand, the bleeding caused by a rupture can have fatal consequences.
On the other hand, the treatment procedure itself is risky and
can lead to severe complications. Especially in the case of small,
asymptomatic cerebral aneurysms, the mortality rate of the treat-
ment may exceed the risk of rupture [Wie03].
Therefore it is vital to assess its risk of rupture to devise an opti-
mal, patient-specific treatment plan. This is especially true for pa-
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
tients with multiple aneurysms that may require several treatment
sessions. Each aneurysm has its own individual risk of rupture,
and each separate treatment session increases the overall procedu-
ral risk. To minimize the risk of both treatment and incidence of
a rupture, the physicians need to identify and treat the aneurysms
with the highest risk of rupture and keep the rest under observation.
2. Related Works
There are various ways to enhance the perception of both depth
and shape in computer-generated 3D images [BCFW08,PBC16].
In this paper, we focus on the use of color to increase the perception
of depth.
Rheingans and Ebert used distance color blending, a combina-
tion of intensity depth cueing and color modulation, to increase
depth perception in volume models [RE01]. This approach mim-
ics the light-scattering effect of the atmosphere by reducing color
intensity of more distant objects and adding a slightly blue tint to
them. Joshi et al. later validated this method specifically for en-
hancing depth perception in vessel visualization [JQD08].
Another method to convey depth by emulating real-world op-
tical effects is depth of field (DoF), where objects are gradually
blurred depending on their distance to a focal plane. Without using
eye tracking, this focal plane needs to be positioned manually or us-
ing heuristics [RSH06]. Grosset et al. evaluated the effectiveness of
various DoF techniques in a study with 25 participants [GSBH13].
They found that DoF only supports depth perception when the fo-
cal plane is placed in the front of the scene. A general problem of
DoF is that it is not possible to focus on two objects at the same
time unless they have a similar distance to the viewer.
Ritter et al. employed hatching to visualize depth relations in
complex vascular structures [RHD06]. Whenever two sections of
the anatomy were overlapping, the posterior structure was hatched
to simulate a shadow. The size of the hatched area directly cor-
responded to the distance between the two structures. In a study
with 160 participants, Ritter et al. were able to show that their ap-
proach significantly increased depth perception when compared to
Gouraud shading. Lawonn et al. presented a combination of depth-
dependent halos, support lines and the illustrative shadows by Rit-
ter et al. to improve perceptibility of depth [LLPH15]. These sup-
port lines are cast from manually selected points of the vessel onto
a plane, creating an effect similar to beams holding the model up.
They could successfully convey the depth of a complex 3D model
in a static 2D image, although it does not allow for a free rotation
of the vessel.
Pseudo Chromadepth (PCD) was introduced by Ropinski et al.
[RSH06]. It is based on the idea of enhancing depth perception in
3D visualizations of angiography datasets by mapping the depth of
each point on the surface to a color gradient. PCD was derived from
the chromadepth technique [Ste87], which follows a similar idea.
Due to the fact that light with different wavelengths is refracted at
different angles in the lens of the eye, color can be used to create the
illusion of depth in an otherwise flat image. This does not necessar-
ily require any special type of surface, glasses or other additional
devices, although the effect can be strongly enhanced by diffraction
grating glasses [BC98].
Figure 1: Comparison of the chromadepth (left) and pseudo chro-
madepth (right) color scales applied to a cerebral vessel.
Instead of using the full range of colors visible to the human eye,
Pseudo Chromadepth only uses a gradient from red (low depth) to
blue (high depth). A wide range of hues might distract from the
shading used to convey shape. The chroma depth color scale may
work well for geometric objects or shapes with low complexity,
such as an organ surface, but it is inappropriate for such complex
shapes as vessels. A comparison between both techniques can be
seen in Figure 1.
The colors red and blue were chosen due to their high difference
in wavelength to maximize their chromadepth effect. Additionally,
red is attention-grabbing and intuitively perceived as foreground,
whereas blue – the color of the sky – is perceived as background.
In a study with 14 participants, Ropinski at al. showed that an-
giography images could benefit from color-encoded depth informa-
tion [RSH06]. Additional studies confirmed this effect [KOCC14].
The Fresnel effect has previously been used to integrate addi-
tional information into vessel visualizations. Gasteiger et al. intro-
duced Ghosted Views, which use an approximation of the Fresnel
effect to modulate the opacity of vessel surfaces [GNKP10]. This
method allows to show the blood flow inside of a vessel without
removing the entire front-facing part of the surface, thus increas-
ing shape perception of the vessel. In a subsequent study, Baer et
al. showed that this approach allowed for a more accurate analysis
of the aneurysm and its flow patterns [BGCP11]. Glaßer et al. pre-
sented a similar technique, which uses the Fresnel effect to high-
light vessel boundaries [GSB16]. They also used discrete color
scales to visualize surface attributes on the vessel surface, but did
not combine this directly with their boundary-enhanced view.
3. Method
The idea of both chromadepth and PCD is based on the fact that
the color channel of the image does not contain any relevant infor-
mation and can therefore be fully utilized to increase depth percep-
tion [RSH06]. While this is true for angiography images, it cannot
be generalized for any kind of medical visualization task.
When analyzing vascular pathologies, physicians are often inter-
ested not only in the vessel shape, but also in functional parameters,
such as pressure or wall shear stress on the vessel wall. An appro-
priate visualization should therefore convey the general shape and
depth of the vessel model, but simultaneously encode the aforemen-
tioned functional parameters as well. The physician should be able
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
Figure 2: Comparison of a smooth color scale (A), discrete color
scale (B) and discrete color scale with additional boundaries (C)
when visualizing wall shear stress on a vessel.
to compare different regions on the vessel wall regarding their spa-
tial relation and parameter values without having to switch between
different types of shading.
These requirements prevent the application of traditional chro-
madepth or PCD shading, as it would conflict with the parameter
information encoded into the color channel. A likely result would
be a slower analysis with increased risk of errors and mental load
for the physician. In contrast, we present a technique that allows
the use of PCD in addition to mapping data to the surface color of
a model, which is described in the following.
3.1. Surface Visualization
The vessel surface models are generated from 3D digital sub-
traction angiography data with cerebral aneurysms by applying
a threshold-based segmentation. The iso-surface is extracted and
converted into a triangle mesh. This mesh is then visualized as a
3D surface model and illuminated using Phong Shading with a sin-
gle headlight.
For the extraction of hemodynamic parameters, such as pressure
and wall shear stress, the surface mesh is employed for computa-
tional fluid dynamic simulations. We realized two approaches: the
parameters are mapped to a color scale ranging from white to or-
ange for a pilot study and to a color scale from white to green for
the final study. We decided against using hatching to convey the ad-
ditional information, as it may also interfere with the Phong-based
lighting we use to convey the shape of the vessels. Furthermore,
hatching is not well suited to highlight small areas of interest.
When analyzing surface parameters on a vessel, physicians often
look for "hot spots". These are small areas with very high values,
which can be perceived pre-attentively when they are encoded with
color. To highlight regions with particularly high or low parameter
values, we have chosen to discretize the color scale to five different
shades (Fig. 2B). To make these shades even more distinct, a black
outline has been added to mark the transition line between shades
(Fig. 2C).
3.2. Fresnel-Inspired PCD
Traditionally, PCD occupies the entire color channel. This mostly
prevents additional information from being shown on the objects
surface. Since the green color channel is not occupied by PCD, a
trivial solution would be to map information to that specific color
Figure 3: PCD shading where the depth is continuously mapped
to the red and blue color scale and the scalar parameter is mapped
to the green color channel using a discrete scale.
channel only. Such a visualization (Fig. 3) would be unsatisfactory,
since it is very difficult for humans to mentally disassemble a color
into their respective channels. The same parameter value can have
widely different hues depending on its location on the model’s sur-
face. As a result, the interpretation of such a visualization is very
challenging.
Our method displays the PCD color scale on the edges of the
3D model only, based on the current viewing direction. This type
of shading is inspired by the Fresnel effect, which describes the
amount of reflection and refraction of light on a surface in relation
to the viewing angle. A flatter viewing angle on a surface increases
the amount of light that is reflected, resulting in the surface appear-
ing brighter when lighted (Fig. 4A).
Steep angle
Weak reflection
Flat angle
Strong reflection
Weak reflection
Strong reflection
Strong reflection
on the edges
on the edges
in the center
A
B
Figure 4: Principle of the Fresnel effect; the amount of reflection
on a reflective surface depends on the viewing angle (A). When
applied to spherical object; the edges exhibit strong reflections due
to the shallow viewing angle (B).
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
A physically accurate calculation of this effect is quite compli-
cated, especially when taking into consideration that due to chro-
matic dispersion, the strength of the Fresnel effect also depends
on the light components’ wavelengths. Instead, we use a simplified
version of this effect to generate a mask for overlaying the PCD
color gradient. Our Fresnel-Inspired PCD (FI-PCD) mask MPCD is
calculated similarly to ghosted views [GNKP10] using the follow-
ing formula:
MPCD =fscale (1− |2arccos(
~
I·~
N)
π1|)s
~
Iand ~
Nare the incident and normal vectors at the surface re-
spectively. The scaling factor fscale can be used to adjust the ef-
fect strength. Similarly, the variable scontrols the steepness of the
transition from surface to PCD scale. In our application, we have
empirically chosen a scaling factor of 1 and a steepness of 2.
MPCD is dependent on the angle between the normal and inci-
dent vectors, reaching its maximal value when they are orthogonal
to each other. On spherical or tubular models, the Fresnel effect
strongly increases the reflectiveness around the edges of the model
(Fig. 4B).
Our final FI-PCD visualization comprises two different images,
both of them renderings of the vessel surface. The first one has
the parameters mapped to its color (Fig. 5A), the second is colored
entirely according to the PCD scale (Fig. 5B). For each pixel in the
final image, the pixel’s value in the mask MPCD is extracted and
used as weight for the linear interpolation between the two images
(Fig. 5C). For example, black MPCD pixels yield the color-coded
parameter value and white MPCD pixels yield the PCD-based color-
coding.
The resulting FI-PCD visualization (Fig. 5D) allows mapping
a scalar parameter to any color scale, while PCD depth cues are
shown only on the edges of the model. They are still clearly visible
to the user while interference with the object’s surface color is re-
duced. Usually, the physician would rotate the vessel in a way that
the interesting areas are facing the camera instead of being rele-
gated to the edge of the model. In addition to providing depth cues
by hue, displaying the PCD scale at the edges also increases the
perceptibility of overlaps, which is another important depth cue.
3.3. Implementation
Generation and composition of both images is performed mostly in
the fragment shader.
When rendering the surface, the attribute values for the surface
are sent to the graphics card as vertex attributes. Additionally, the
highest and lowest values of the attribute as well as the highest and
lowest depth values from the previously rendered frame are stored
in the fragment shader as uniform variables. Then, the attributes
are interpolated between vertices, normalized to a [0,1]range and
transformed into a color value by the fragment shader. The trans-
formation is performed linearly in RGB color space between white
(#ffffff ) and orange (#ff7f00, pilot study) or green (#00ff00, final
study). Next, the resulting color value is discretized into five dis-
tinct shades and used as surface color.
Figure 5: Composition of images to create the FI-PCD visualiza-
tion: Surface color image (A), PCD image (B), Composition mask
(C) and resulting FI-PCD image (D).
The boundaries between color shades are generated dynamically
on a per-triangle base by analyzing the affinity of each vertex to
a certain color class. For each triangle with different affinities at
the edges, the fragment shader draws a black line separating these
vertices. This approach allows for a very fast generation of dynamic
outlines on the surface, without the need for any pre-processing
or the creation of new geometry. Unfortunately, since the lines are
always at the center between two vertices, they do not always line
up exactly with the actual color transition. On a model with a decent
triangle resolution, this effect is only noticeable when zooming in
very closely to the surface.
The second image is generated by normalizing the current frag-
ment’s depth using the previously stored depth range and mapping
the resulting value to the PCD color scale. Using the depth range
from the previous frame allows us to draw the geometry using
a single rendering pass, although it produces a barely noticeable
flicker in the PCD color scale during fast animations. Afterwards,
the MPCD value is calculated and used to compute the composition
of both images.
4. Evaluation
Our evaluation consists of two separate studies; a pilot study with
participants from the general public, and a final study with experts
in the fields of medical engineering and flow simulation. All par-
ticipants were shown 3D visualizations of intra-cranial vessel sur-
faces models. These models had one of multiple available attributes
mapped to their surface, such as pressure or wall shear stress. Par-
ticipants were shown two points on these datasets and had to select
either the one closest to them or the one with the highest parameter.
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
Figure 6: The different shadings used in the first study: No depth
cues, trivial brightness cues and FI-PCD (f.l.t.r.)
The datasets were shown with three different shading styles. The
first style was a normal, phong-shaded visualization without any
distinct depth enhancement. The second used brightness as a depth
cue. Distant triangles were reduced in brightness, with the highest
possible reduction being 75%. This value was chosen empirically
as a trade-off between having a strong effect on depth perception
while still being able to discern the color of farther away parts of
the model. The last style was our implementation of FI-PCD. All
three visualization styles can be seen in Figure 6.
We expected the visualization without depth cues to perform
worst in the depth judgment, but best in the parameter judgment
task. As both brightness-based cues and FI-PCD would partially
overlay the surface color scale, we expected them to perform
equally well, but not as good as the visualization without depth
cues. Since PCD has proven superior to brightness- or contrast-
based cues by studies in the past (such as [KOCC14]), we expected
FI-PCD to perform best in the depth judgment task.
The pilot study allowed us to identify several flaws in our tech-
nique. Before the final study, we corrected these problems by
changing some aspects of both our visualization as well as the
application. First, the color scale used to encode the surface at-
tribute was changed from white-to-orange to white-to-green. The
original orange scale was chosen due to aesthetic reasons. How-
ever, many participants in the pilot study noted that red areas from
the FI-PCD shading interfered with orange areas from the surface
attribute color scale. Since PCD only uses the red and blue color
channel, green was chosen for the surface attribute to prevent color
overlaps. A comparison between the two color scales in combina-
tion with FI-PCD can be seen in Figure 7.
We also added a permanent legend for the used color scales in
the bottom left corner of the screen. This was done in response to
some participants in the first study confusing the meaning of some
of the colors during the course of the study. The legend always
encoded the surface attribute color scale in combination with the
current depth enhancement color scale. All three scales can be seen
in Figure 8.
4.1. Application
When started, the test application presents the user with a few in-
structional pages. All of them include a "Continue"-button that be-
comes enabled after five seconds and allows the participant to ad-
Figure 7: Different color scales used in the first (left) and second
(right) study in combination with FI-PCD.
vance to the next screen. The first and second pages contain general
information about the study as well as labeled example images for
all types of visualizations used in this study, as seen in Figure 9. To
prevent any bias, these images are always shown in a random order.
The actual study consists of two blocks, where the user has to
select either the point closest to them, or the one with the high-
est scalar surface parameter. Therefore, the user has to identify the
spatial relation or ranking of scalar values of two selected points
on the surface (Fig. 10). Additionally, they always have the option
to click a button labeled "Not sure" if they cannot decide for one
of the points. During each task, the application measures the com-
pletion time, rotation time and whether the user clicked the correct
point or hit the "Not sure"-button instead. For the rotation time, we
counted the amounts of single frames that a rotation was performed
in and converted them to a duration in seconds. Frames where the
user kept the left mouse button pressed without moving the mouse
(therefore not actually performing a rotation) were not included.
Each block is introduced by another instructional page, which
is then followed by six dedicated tutorial datasets. They serve as
a way to familiarize the user with the visualizations and tasks,
therefore their measurements are excluded from the final statistic.
Furthermore, a learning effect during the actual evaluation is pre-
vented.
After completing the tutorial for each block, the user sees a mes-
sage explaining that the training part is over and asking them if they
have any questions before proceeding. This was done to ensure they
were properly prepared and did not have to ask questions during
the time-measured evaluation. They were encouraged to complete
Figure 8: The color scale legends shown in the final study: No
depth cues, trivial brightness cures and FI-PCD (f.l.t.r.)
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
Figure 9: Instructional page detailing the different visualizations
(ordered randomly). For the second study, the images were updated
to reflect the changed surface color scale.
each task as fast and accurately as possible due to the time mea-
surement. "Guessing" the correct answer was discouraged in favor
of using the "Not sure"-button.
The order of blocks was consistent for all participants, starting
with the depth judgment tasks and then switching over to the pa-
rameter judgment task. Each task consisted of 30 images in total,
six of them being the training images. All users were shown the
same images, although they were ordered randomly. The applica-
tion ensured that the same dataset did not appear twice in a row.
The participants did not receive any immediate feedback about the
correctness of their answers during the study, but statistics about
their general performance were made available to them afterwards
upon request.
4.2. Questionnaire
After completing the assignments on the computer, all participants
were asked to fill out a questionnaire. In addition to age, biologi-
cal gender and known visual disorders, participants were asked if
they have experience with analyzing medical data or modeling 3D
Figure 10: One of the datasets with two marked points shown to
the participants as part of the depth judgment task. The image used
brightness-based depth cues and the green color scale from the fi-
nal study.
Figure 11: Experimental setup for the pilot study. The two disabled
monitors in the corner were not part of the study.
objects and whether they play 3D video games regularly. For vi-
sual disorders, we were mostly interested in those that impede the
ability to perceive color or depth. Since there are many cases where
people are unaware of their color perception impairment, we added
a very abbreviated color blindness test using three Ishihara plates.
Two of them had numbers encoded in them (42 and 6) while the last
one did not. None of the participants that had not already denoted
a form of color blindness in the questionnaire failed this test.
At the end, participants were asked to rank the three types of vi-
sualizations according to their usefulness for perceiving depth and
the surface attribute as well as their general aesthetic. They were
also given space for any additional remarks.
5. Pilot Study
For the pilot study, we took advantage of the popular open house
day at our university as a means of finding volunteers. Visitors of
this event were asked to participate in our study. Both verbal ex-
planations as well as written instructions and other materials were
made available to the participants.
Half of the participants were randomly selected to be given lim-
ited control over the camera during the study, whereas they can ro-
tate the dataset by ten degrees in any direction. These participants
were shown an extra paragraph in one of the instructional pages
of the application explaining that they had the ability to orbit the
camera. If they did not rotate the camera at least once during the
tutorial, they were reminded by a pop-up dialog.
5.1. Participants
A total number of 104 people from the general public volunteered
to participate in the pilot study. Ten of them were later rejected
due to vision impairments (i.e. various forms of color blindness or
problems with depth perception), failing to comprehend the assign-
ment or not filling out the corresponding questionnaire. The age
of the participants ranges from 11 to 73, with an average of 28.6
and a standard deviation of 14. Out of the 94 participants that were
included in the evaluation, 40 were female (42.6 %).
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
5.2. Setup
In order to allow for a high number of participants, we set up four
PCs for simultaneous use (Fig. 11). They were positioned in a cor-
ner of the room to prevent distractions from the rest of the event.
To ensure comparability of the results between the different sta-
tions, we used PCs with similar hardware specifications and identi-
cal screens. All stations ran the application at a resolution of 1920
×1080 with 60 frames per second.
The participants were given a short verbal introduction about the
topic of vessel visualization in general and the study in particular.
To keep any descriptions simple and explanations short, the differ-
ent scalar attributes shown in the visualizations were always just
referred to as "pressure" despite also including wall shear stress.
After that, they were instructed to sit down at one of the stations
and follow the on-screen instructions from the application.
5.3. Results
For each participant, we calculated four values from our measure-
ment for each shading style. Correctness is the percentage of cor-
rect answers, e.g. how often participants selected the nearest point
(during the depth judgment task) or the point with higher scalar
value (during the parameter judgment task), respectively. Certainty
denotes the percentage of answers where the user selected any of
the points and not the "Not sure"-button. Duration is the average
time in seconds the users took for each image. Rotation is the aver-
age time the user spend rotating the dataset. For this value, we only
included users who actually rotated the dataset.
The ability of the participants to pick the point closest to them
benefited from having any form of depth cues enabled (Fig. 12).
Without them, they were only correct in 79 % of the depth judg-
ment tasks. Brightness-based depth cues increased their accuracy
to 90 %, whereas FI-PCD only increased it to 85%. This is surpris-
ing, as we were expecting the FI-PCD to provide much better depth
cues than the brightness-based approach.
Although being reminded after each training session that they
could rotate, only 35 of the 50 users with the ability to rotate ac-
tually made use of it. Three of them performed so little rotation
that we assume that to be accidental. This was possibly a result
of being overwhelmed due to unfamiliarity with 3D visualizations.
Users that stated experience in 3D modeling or 3D video games ro-
tated for an average of 0.3 seconds per dataset, whereas users with
no experience only rotated for 0.19 seconds.
The values for certainty and rotation are extremely similar for
each of the three shading styles. The users generally rarely used the
"Not sure"-button in this study. The average duration was slightly
higher for the visualization without depth cues (4.2 s) in compari-
son to brightness-based cues (3.8 s) and FI-PCD (4.0 s).
For the parameter judgment task, the visualization without cues
reached the best average correctness (96 %, Fig. 13). Users also
performed fastest, with an average duration of 2.9 s per image. This
result was to be expected, as there are no additional color or bright-
ness gradients added to the surface color. The brightness-based
depth cues performed better than FI-PCD in regards to correctness
(92 % compared to 80 %) and duration (3.2 s compared to 3.5 s).
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
ݱ®®»½¬²»
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õ
õõ
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
Ý»®¬¿·²¬§
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õ
õ
õ
ð
î
ì
ê
è
ïð
ïî
ïì
Ü«®¿¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
ðòð
ðòë
ïòð
ïòë
îòð
îòë
ન¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
Figure 12: Box plots showing correctness (top left), certainty (top
right), duration in seconds (bottom left) and rotation duration in
seconds (bottom right) for the depth judgment task over all partic-
ipants from the pilot study.
This most likely stems from our choice of color scale to encode the
surface parameters in this study. Many participants remarked that
the orange from the surface color scale was interfering with the red
from the PCD scale, thus making it hard to distinguish them.
Interestingly, in order to interpret the colors of areas strongly
affected by depth cues (i.e. those close to the edge of the vessel
when using FI-PCD or those in the background when using bright-
ness cues), users often resorted to "counting" color gradients. They
would search for an area that was completely white and then count
the boundaries they had to cross to reach the marked point. That
way, they could tell which area represented a higher parameter
value even without being able to distinguish the colors directly.
Since this approach requires a visually uninterrupted path from a
marked point to a white area, it was not possible in all cases.
Just as in the depth judgment task, the certainty for all shading
styles is very similar. Rotation was used even more rarely in this
task. Since the marked points were never obstructed by other ge-
ometry, there was little point in rotating the dataset to compare the
surface coloring.
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
ݱ®®»½¬²»
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõ
õ
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
Ý»®¬¿·²¬§
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õ
õõ
ð
î
ì
ê
è
ïð
Ü«®¿¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
ðòð
ðòë
ïòð
ïòë
ન¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
Figure 13: Box plots showing correctness (top left), certainty (top
right), duration in seconds (bottom left) and rotation duration in
seconds (bottom right) for the parameter judgment task over all
participants from the pilot study.
We also analyzed the correctness in regards to whether the users
made use of rotation during the tasks. The ability to rotate the view
had very little effect on the results of the surface parameter task
(Fig. 14, bottom). The correctness of the depth judgment task in-
creased when rotation was used on the FI-PCD images as well as
those without depth cues. Since parallax movement is another im-
portant depth cue, this improvement is not surprising.
6. Final Study
For the final study, we directly approached several experts in the
fields of medical engineering and flow simulation. Due to the lower
number of participants in this study, we decided against splitting
them into two groups. Therefore, we allowed all of them to rotate
the camera.
6.1. Participants
Eleven experts volunteered to take part in our final study. One per-
son was excluded due to color blindness. The age of the included
Figure 14: Influence of rotation on the ability to judge depth (top)
and surface parameters (bottom)
participants ranged between 22 and 41 (average of 29.1), with two
of them being female (20 %).
6.2. Setup
The second study was performed on a laptop, as it took place at
our participant’s workplace. Despite having less powerful hardware
than the PCs used in the first study, it was still able to run the ap-
plication at 1920 ×1080 with 60 frames per second.
For this study, the introduction to vessel visualization was either
omitted or kept very brief, since most participants were familiar
with this field already. The instructions given by the application
itself remained unchanged from the pilot study.
6.3. Results
In our second study, the FI-PCD method reached better results (Fig.
15). During the depth judgment task, users were able to pick the
correct point in 94 % of the cases. With the brightness-based shad-
ing, they were able to choose correctly in 90 % of the cases. Without
any depth cues, the participants only reached 85 % accuracy.
The same trend is visible in the certainty plots. Overall, the dura-
tion and rotation plots from the second study show the same trends
as those in the first study. Interestingly, users took longer for their
decision and also rotated the view more when viewing the datasets
with FI-PCD compared to brightness-based depth cues. This may
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
be due to the fact that the combination of PCD and surface color
scale can no longer be pre-attentively perceived.
As expected, users were able to judge the parameters best when
no depth cues were present, reaching a mean correctness of 98 %.
Brightness-based depth cues produced an almost identical result
with a mean correctness of 96 %. FI-PCD shading had the strongest
negative effect on the participant’s ability to compare parameter
values on the surface, although not as strong as in the first study.
The mean correctness in this case was 90 %.
The average duration for each decision (from both tasks) was
significantly higher in the second study (5.5 s) compared to the
first (3.7 s). Similarly, the average rotation duration was also higher
(0.3s compared to 0.2 s). This may indicate that in the second study,
participants put more effort into the evaluation.
7. Discussion
Our studies have shown that FI-PCD can increase the perception
of depth while maintaining recognizability of surface scales on
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
ݱ®®»½¬²»
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
Ý»®¬¿·²¬§
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
ð
î
ì
ê
è
ïð
ïî
ïì
Ü«®¿¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õ
õõ
ðòð
ðòë
ïòð
ïòë
ન¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õ
õõ
Figure 15: Box plots showing correctness (top left), certainty (top
right), duration in seconds (bottom left) and rotation duration in
seconds (bottom right) for the depth judgment task over all partic-
ipants from the final study.
the vessel surface. For the latter, a careful choice of color scale
is required to avoid conflicts with the color gradients introduced
by PCD. In our first study, we used an inappropriate color scale
to encode surface parameters. This strongly reduced our method’s
ability to convey both depth and surface parameters. We were able
to remedy this problem in the second study by choosing a different
scale that relies only on the green color channel, which goes unused
by PCD. This resulted in a higher increase of depth perception than
classic, brightness-based depth cues.
We decided to use a discretized color scale instead of a smooth
one. This reduces ambiguity between the surface color and PCD
scale while at the same time highlights areas with high or low val-
ues, which physicians are often interested in since their decisions
are discrete as well. The highlighting was increased further by the
introduction of outlines around the differently colored surface re-
gions. This created a robust visualization that still allowed users to
compare parameter values on the surface even when overlaid with
another color or brightness gradient.
Both studies showed that overlaying the color channel with depth
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
ݱ®®»½¬²»
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
Ý»®¬¿·²¬§
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
ð
î
ì
ê
è
ïð
Ü«®¿¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õõõ
ðòð
ðòî
ðòì
ðòê
ðòè
ïòð
ન¬·±²
Ò± ½«» Þ®·¹¸¬²» Ú×-ÐÝÜ
õ
õõ
Figure 16: Box plots showing correctness (top left), certainty (top
right), duration in seconds (bottom left) and rotation duration in
seconds (bottom right) for the parameter judgment task over all
participants from the final study.
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
B. Behrendt & P. Berg & B. Preim & S. Saalfeld / Combining PCD and Parameter Mapping for Vascular Surface Models
cues reduces the recognizability of the surface color scale. This ef-
fect was strongest when using FI-PCD. A likely explanation is that
FI-PCD affects the color of both close and distant regions, whereas
brightness-based depth cues only affect distant regions. Therefore,
FI-PCD should be kept as an optional addition to any visualization
that can be disabled in case an in-depth comparison between the
scalar values of different surface points is required.
8. Conclusion & Future Work
With FI-PCD, we have introduced a novel rendering technique
that combines Pseudo Chromadepth with color-encoded surface at-
tributes to visualize vascular anatomy in combination with scalar
parameters. We have performed two studies to evaluate our tech-
nique. While we could show that FI-PCD can enhance the percep-
tion of depth, there are still issues that need to be improved on.
First, FI-PCD tends to distort the underlying color scale. This
can be partly remedied by choosing a scale that does not interfere
with the red and blue colors from PCD, such as our white-to-green
scale. However, it would be interesting to see if this effect can be
further reduced by using different values for the scaling factor and
steepness in our FI-PCD formula. Reducing the scaling factor or in-
creasing the steepness would make the PCD color scale less promi-
nent in the visualization. Therefore, it may be possible to find a
setting that results in a better trade-off between depth and surface
color perception.
In many scenarios, the physician would not only be interested
in a surface parameter, but also blood flow patterns. Simply dis-
playing them inside of the vessel anatomy using established smart
visibility techniques would likely produce unsatisfactory results. In
addition to having to cut or fade away parts of the surface to re-
veal the underlying flow (thus making it harder to see the surface
color), displaying path lines with their own color scale would also
add another layer of complexity to the color channel. A seamless
way of integrating flow visualization would therefore be a useful
extension.
Acknowledgments
We would like to warmly thank Prof. Douglas W. Cunningham and
Maria Lutz for their guidance and helpful remarks concerning the
study design. Thanks also go to Patrick Saalfeld for his help with
the statistical evaluation of our study results.
The work of this paper is funded by the European Regional De-
velopment Fund under the operation number ’ZS /2016/04/78123’
as part of the initiative "Sachsen-Anhalt WISSENSCHAFT Schw-
erpunkte".
References
[BC98] BAILEY M. , CLA RK D.: Using ChromaDepth to Obtain Inex-
pensive Single-image Stereovision for Scientific Visualization. Journal
of Graphics Tools 3, 3 (1998), 1–9. 2
[BCFW08] BART Z D., C UNNINGHAM D. W., FISC HE R J., WAL L-
RAVEN C .: The Role of Perception for Computer Graphics. In Euro-
graphics (STARs) (2008), pp. 59–80. 2
[BGCP11] BAE R A., G AS TEI GER R ., CUNNINGHAM D., P REI M B.:
Perceptual evaluation of ghosted view techniques for the exploration of
vascular structures and embedded flow. In Computer Graphics Forum
(2011), vol. 30, pp. 811–820. 2
[BSB06] BONNEVILLE F., S OUR OUR N ., BIONDI A .: Intracranial
aneurysms: An overview. Neuroimaging clinics of North America 16,
3 (2006), 371–82, vii. 1
[GNKP10] GA STE IGE R R., NEU GEB AUER M ., KUBISCH C., P REI M
B.: Adapted Surface Visualization of Cerebral Aneurysms with Em-
bedded Blood Flow Information. In VCBM (2010), pp. 25–32. 2,4
[GSB16] GL ASSE R S., S AA LFE LD P., BE RG P., ME RTE N N., P REI M
B.: How to Evaluate Medical Visualizations on the Example of 3D
Aneurysm Surfaces. In Eurographics Workshop on Visual Computing
for Biology and Medicine (2016). 2
[GSBH13] GRO SSE T A., SCH OTT M., BO NNE AU G.-P., HAN SE N
C. D.: Evaluation of Depth of Field for depth perception in DVR. In
2013 IEEE Pacific Visualization Symposium (PacificVis) (2013), IEEE,
pp. 81–88. 2
[JQD08] JO SHI A ., QIAN X ., DIONE D. P., BU LSA RA K. R ., BREU ER
C. K., SINUSAS A . J., PAPADEMETRIS X.: Effective visualization
of complex vascular structures using a non-parametric vessel detection
method. IEEE transactions on visualization and computer graphics 14,
6 (2008), 1603–1610. 2
[KOCC14] KER STE N-OERT EL M. , CH EN S. J.-S., COL LIN S D. L.: An
evaluation of depth enhancing perceptual cues for vascular volume visu-
alization in neurosurgery. IEEE transactions on visualization and com-
puter graphics 20, 3 (2014), 391–403. 2,5
[LLPH15] LAWO NN K. , LU Z M., P REI M B., HAN SEN C .: Illustrative
Visualization of Vascular Models for Static 2D Representations. In Med-
ical Image Computing and Computer-Assisted Intervention – MICCAI
2015. 2015, pp. 399–406. 2
[MKH12] MO RITA A. , KIRINO T., H AS HI K. , AOKI N., F UKUHARA
S., HAS HIM OTO N., NAK AYAMA T., SAKA I M., T ER AMOT O A.,
TOMINARI S., YOSH IMO TO T.: The natural course of unruptured cere-
bral aneurysms in a Japanese cohort. The New England journal of
medicine 366, 26 (2012), 2474–2482. 1
[PBC16] PR EIM B ., BA ER A. , CUNNINGHAM D. , ISENBERG T.,
ROPINSKI T.: A survey of perceptually motivated 3d visualization of
medical image data. In Computer Graphics Forum (2016), vol. 35,
pp. 501–525. 1,2
[RE01] RHEINGANS P., EB ERT D.: Volume illustration: Nonphotoreal-
istic rendering of volume models. IEEE transactions on visualization
and computer graphics 7, 3 (2001), 253–264. 2
[RHD06] RI TTE R F., HA NSE N C., D ICKEN V., KONRAD O., PRE IM
B., PEI TGE N H.-O.: Real-time illustration of vascular structures. IEEE
transactions on visualization and computer graphics 12, 5 (2006), 877–
884. 2
[RSH06] ROPINSKI T., STEINICKE F., HINRICHS K.: Visually Sup-
porting Depth Perception in Angiography Imaging. In Smart Graphics,
vol. 4073 of Lecture Notes in Computer Science. 2006, pp. 93–104. 2
[Ste87] ST EEN BLI K R. A.: The Chromostereoscopic Process: A Novel
Single Image Stereoscopic Process. McAllister D. F., Robbins W. E.,
(Eds.), SPIE Proceedings, pp. 27–34. 2
[Wie03] WIE BER S D. O .: Unruptured intracranial aneurysms: Natural
history, clinical outcome, and risks of surgical and endovascular treat-
ment. The Lancet 362, 9378 (2003), 103–110. 1
c
2017 The Author(s)
Eurographics Proceedings c
2017 The Eurographics Association.
... Applying pseudo chroma-depth to a whole mesh, however, impairs the ability to use other shading techniques to convey structure or to display additional information on the surface itself. To cope with this issue, Behrendt et al. [BBPS17] proposed to apply pseudo chroma-depth only to the contour region of a mesh to make space for supplementary data. Apart from colouring or texturing a given surface to convey data, additional geometry can be added as glyphs to an existing scene [ROP11]. ...
... The smaller the diameter of a vessel segment, the better can information from the contour represent the data of the affiliated vessel segment. For example, this has previously been done by Behrendt et al. [BBPS17], who encoded the depth of vasculature by colouring its contour using pseudo chroma-depth. Instead of defining the contour on the surface itself (e.g. by using a Fresnel approximation), we aim to create additional geometry at locations of the rendered mesh, where the surface normal is orthogonal to the view direction. ...
... Alternatively, the contour can simply be coloured w.r.t. the pseudo chroma-depth spectrum as in Figure 15, leaving the surface free for other encodings. This addresses an issue which has been tackled by Behrendt et al. [BBPS17]. They mixed the colour coding of a scalar field and pseudo chroma-depth using a Fresnel term. ...
Article
Full-text available
The study of vascular structures, using medical 3D models, is an active field of research. Illustrative visualizations have been applied to this domain in multiple ways. Researchers made the geometric properties of vasculature more comprehensive and augmented the surface with representations of multivariate clinical data. Techniques that head beyond the application of colour‐maps or simple shading approaches require a surface parameterization, that is, texture coordinates, in order to overcome locality. When extracting 3D models, the computation of texture coordinates on the mesh is not always part of the data processing pipeline. We combine existing techniques to a simple parameterization approach that is suitable for tree‐like structures. The parameterization is done w.r.t. to a pre‐defined source vertex. For this, we present an automatic algorithm, that detects the tree root. The parameterization is partly done in screen‐space and recomputed per frame. However, the screen‐space computation comes with positive features that are not present in object‐space approaches. We show how the resulting texture coordinates can be used for varying hatching, contour parameterization, display of decals, as additional depth cues and feature extraction. A further post‐processing step based on parameterization allows for a segmentation of the structure and visualization of its tree topology. The study of vascular structures, using medical 3D models, is an active field of research. Illustrative visualizations have been applied to this domain in multiple ways. Researchers made the geometric properties of vasculature more comprehensive and augmented the surface with representations of multivariate clinical data. Techniques that head beyond the application of colour‐maps or simple shading approaches require a surface parameterization, that is, texture coordinates, in order to overcome locality. When extracting 3D models, the computation of texture coordinates on the mesh is not always part of the data processing pipeline.
... Such information contains for instance additionally acquired or derived parameters such as blood flow, pressure or wall shear stress, which are typically conveyed by stream lines, different color scales or glyphs. Effectively communicating all this information without overwhelming the viewer is a problem that has been addressed by several authors in the past [2,17,19,30]. ...
... Thus, various approaches have been considered to overcome this problem. Behrendt et al. [2] introduced a technique to encode information on the vessels whilst maintaining the benefits of the pseudo-chromadepth color scale, by applying it on the edges of the vessels only. Lichtenberg et al. [19] used glyphs to communicate depth information on the vessel end-points freeing the surface of the vessel of such task. ...
... Domain experts are often interested in visualizing also other properties on top of the vessels. This problem was addressed by Behrendt et al. [2] who combined pseudo-chromadepth with additional information on top of the vessels. They used the pseudo-chromadepth color scheme to shade the areas close to the contour by applying a blending mask inspired by Fresnel equations. ...
Preprint
To enhance depth perception and thus data comprehension, additional depth cues are often used in 3D visualizations of complex vascular structures. Accordingly, there is a variety of different approaches described in the literature, ranging from chromadepth color coding over depth of field to glyph-based encodings. Unfortunately, the majority of existing approaches suffers from the same problem. As these cues are directly applied to the geometry's surface, the display of additional information, such as other modalities or derived attributes, associated with a vessel is impaired. To overcome this limitation we propose Void Space Surfaces which utilize the empty space in between vessel branches to communicate depth and their relative positioning. This allows us to enhance the depth perception of vascular structures without interfering with the spatial data and potentially superimposed parameter information. Within this paper we introduce Void Space Surfaces, describe their technical realization, and show their application to various vessel trees. Moreover, we report the outcome of a user study which we have conducted in order to evaluate the perceptual impact of Void Space Surfaces as compared to existing vessel visualization techniques.
... Applying chromadepth to a 3D surface makes it difficult to additionally encode attributes on the surface. Therefore, Behrendt et al. [22] used the Fresnel term to combine chromadepth and additional parameters. Illustrative techniques were also used to improve depth perception. ...
... In the first step a novel technique is developed. Applying illustrative techniques [7,23,26] , glyphs [27] , or add an additional layer of information [22,28] can improve depth perception, whereas methods aiming at improving shape perception of surfaces use, e.g., Phong shading [34] or line drawings [33] . The last step reveals the potential benefits of the novel technique. ...
... Thus, we combine a 2D and 3D view of the vasculature to highlight clinically relevant properties in a clearer way. For example, the distance of vessel segments to tumor tissue inside a liver, which may aid analysis prior a tumor resection, or hemodynamics data [BBPS17]. A crucial factor that determines what kind of visualization techniques can be applied, is the type of data in which the vasculature is represented. ...
Conference Paper
Full-text available
In this paper, we make contributions to the visualization of vascular structures. Based on skeletal input data, we provide a combined 2D and implicit 3D visualization of vasculature, that is parameterized on-the-fly for illustrative visualization. We use an efficient algorithm that creates a distance field volume from triangles and extend it to handle skeletal tree data. Sphere-tracing this volume allows to visualize the vasculature in a flexible way, without the need to recompute the volume. Illustrative techniques, that have been frequently applied to vascular visualizations often require texture coordinates. Therefore, modifying an object-based algorithm, we propose an image-based, hierarchical optimization process that allows to derive periodic texture coordinates in a frame-coherent way and suits the implicit representation of the vascular structures. In addition to the 3D surface visualization, we propose a simple layout algorithm that applies a 2D parameterization to the skeletal tree nodes. This parameterization can be used to color-code the vasculature or to plot a 2D overview-graph, that highlights the branching topology of the skeleton. We transfer measurements, done in 3D space, to the 2D plot in order to avoid visual clutter and self occlusions in the 3D representation. A visual link between the 3D and 2D views is established via color codes and texture patterns. The potential of our pipeline is shown in several prototypical application scenarios.
... Winne et al., 2011, has used alternatively colored guiding lines for distance visualization [19]. Furthermore, Pseudo Chromadepth is a well known technique for depth perception in 3D angiography datasets, where the color gradient corresponds to the distance value at each point [20], [21]. Although, the technique enhances the relative distance perception between various points in cluttered angiography dataset, it doesn't give a real-time feedback for depth perception which is important during a real surgery. ...
Conference Paper
Augmented reality (AR) is a promising technology where the surgeon can see the medical abnormality in the context of the patient. It makes the anatomy of interest visible to the surgeon which otherwise is not visible. It can result in better surgical precision and therefore, potentially better surgical outcomes and faster recovery times. Despite these benefits, the current AR systems suffer from two major challenges; first, incorrect depth perception and, second, the lack of suitable evaluation systems. Therefore, in the current paper we addressed both of these problems. We proposed a color depth encoding (CDE) technique to estimate the distance between the tumor and the tissue surface using a surgical instrument. We mapped the distance between the tumor and the tissue surface to the blue-red color spectrum. For evaluation and interaction with our AR technique, we propose to use a virtual surgical instrument method using the CAD model of the instrument. The users were asked to reach the judged distance in the surgical field using the virtual tool. Realistic tool movement was simulated by collecting the forward kinematics joint encoder data. The results showed significant improvement in depth estimation, time for task completion and confidence, using our CDE technique with and without stereo versus other two cases, that are, Stereo-No CDE and No Stereo-No CDE.
Article
To enhance depth perception and thus data comprehension, additional depth cues are often used in 3D visualizations of complex vascular structures. There is a variety of different approaches described in the literature, ranging from chromadepth color coding over depth of field to glyph-based encodings. Unfortunately, the majority of existing approaches suffers from the same problem: As these cues are directly applied to the geometry's surface, the display of additional information on the vessel wall, such as other modalities or derived attributes, is impaired. To overcome this limitation we propose Void Space Surfaces which utilizes empty space in between vessel branches to communicate depth and their relative positioning. This allows us to enhance the depth perception of vascular structures without interfering with the spatial data and potentially superimposed parameter information. With this paper, we introduce Void Space Surfaces, describe their technical realization, and show their application to various vessel trees. Moreover, we report the outcome of two user studies which we have conducted in order to evaluate the perceptual impact of Void Space Surfaces compared to existing vessel visualization techniques and discuss expert feedback.
Article
User studies are indispensable for visualization application papers in order to assess the value and limitations of the presented approach. Important aspects are how well depth and shape information can be perceived, as coding of these aspects is essential to enable an understandable representation of complex 3D data. In practice, there is usually little time to perform such studies, and the establishment and conduction of user studies can be labour-intensive. In addition, it can be difficult to reach enough participants to obtain expressive results regarding the quality of different visualization techniques. In this paper, we propose a framework that allows visualization researchers to quickly create task-based user studies on depth and shape perception for different surface visualizations and perform the resulting tasks via a web interface. With our approach, the effort for generating user studies is reduced and at the same time the web-based component allows researchers to attract more participants to their study. We demonstrate our framework by applying shape and depth evaluation tasks to visualizations of various surface representations used in many technical and biomedical applications.
Conference Paper
Full-text available
The study and visualization of vascular structures, using 3D models obtained from medical data, is an active field of research. Illustrative visualizations have been applied to this domain in multiple ways. Researchers have tried to make the geometric properties of vasculature more comprehensive and to augment the surface with representations of multivariate clinical data. Techniques that head beyond the application of color-maps or simple shading approaches require a sort of surface parameterization, i.e., texture coordinates, in order to overcome locality. When extracting 3D models, the computation of texture coordinates on the mesh is not always part of the data processing pipeline. We combine existing techniques to a simple, yet effective, parameterization approach that is suitable for tree-like structures. The parameterization is done w.r.t. to a pre-defined source vertex. For this, we present an automatic algorithm, that detects the root of a tree-structure. The parameterization is partly done in screen-space and recomputed per frame. However, the screen-space computation comes with positive features that are not present in object-space approaches. We show how the resulting texture coordinates can be used for varying hatching, contour parameterization, the display of decals, as an additional depth cue and feature extraction. Source Code can be found at: https://gitlab.uni-koblenz.de/MedVis/PFE_Tree-like_Structures
Conference Paper
Full-text available
Depth assessment of 3D vascular models visualized on 2D displays is often difficult, especially in complex workspace conditions such as in the operating room. To address these limitations, we propose a new visualization technique for 3D vascular models. Our technique is tailored to static monoscopic 2D representations, as they are often used during surgery. To improve depth assessment, we propose a combination of supporting lines, view-aligned quads, and illustrative shadows. In addition, a hatching scheme that uses different line styles depending on a distance measure is applied to encode vascular shape as well as the distance to tumors. The resulting visualization can be displayed on monoscopic 2D monitors and on 2D printouts without the requirement to use color or intensity gradients. A qualitative study with 15 participants and a quantitative study with 50 participants confirm that the proposed visualization technique significantly improves depth assessment of complex 3D vascular models.
Article
Full-text available
Cerebral vascular images obtained through angiography are used by neurosurgeons for diagnosis, surgical planning, and intraoperative guidance. The intricate branching of the vessels and furcations, however, make the task of understanding the spatial three-dimensional layout of these images challenging. In this paper, we present empirical studies on the effect of different perceptual cues (fog, pseudo-chromadepth, kinetic depth, and depicting edges) both individually and in combination on the depth perception of cerebral vascular volumes and compare these to the cue of stereopsis. Two experiments with novices and one experiment with experts were performed. The results with novices showed that the pseudo-chromadepth and fog cues were stronger cues than that of stereopsis. Furthermore, the addition of the stereopsis cue to the other cues did not improve relative depth perception in cerebral vascular volumes. In contrast to novices, the experts also performed well with the edge cue. In terms of both novice and expert subjects, pseudo-chromadepth and fog allow for the best relative depth perception. By using such cues to improve depth perception of cerebral vasculature, we may improve diagnosis, surgical planning, and intraoperative guidance.
Conference Paper
Full-text available
Cerebral aneurysms are a vascular dilatation induced by a pathological change of the vessel wall and often require treatment to avoid rupture. Therefore, it is of main interest, to estimate the risk of rupture, to gain a deeper understanding of aneurysm genesis, and to plan an actual intervention, the surface morphology and the internal blood flow characteristics. Visual exploration is primarily used to understand such complex and variable type of data. Since the blood flow data is strongly influenced by the surrounding vessel morphology both have to be visually combined to efficiently support visual exploration. Since the flow is spatially embedded in the surrounding aneurysm surface, occlusion problems have to be tackled. Thereby, a meaningful visual reduction of the aneurysm surface that still provides morphological hints is necessary. We accomplish this by applying an adapted illustrative rendering style to the aneurysm surface. Our contribution lies in the combination and adaption of several rendering styles, which allow us to reduce the problem of occlusion and avoid most of the disadvantages of the traditional semi-transparent surface rendering, like ambiguities in perception of spatial relationships. In interviews with domain experts, we derived visual requirements. Later, we conducted an initial survey with 40 participants (13 medical experts of them), which leads to further improvements of our approach.
Conference Paper
A novel stereoscopic depth encoding/decoding process has been developed which considerably simplifies the creation and presentation of stereoscopic images in a wide range of display media. The patented chromostereoscopic process is unique because the encoding of depth information is accomplished in a single image. The depth encoded image can be viewed with the unaided eye as a normal two dimensional image. The image attains the appearance of depth, however, when viewed by means of the inexpensive and compact depth decoding passive optical system. The process is compatible with photographic, printed, video, slide projected, computer graphic, and laser generated color images. The range of perceived depth in a given image can be selected by the viewer through the use of "tunable depth" decoding optics, allowing infinite and smooth tuning from exaggerated normal depth through zero depth to exaggerated inverse depth. The process is insensitive to the head position of the viewer. Depth encoding is accomplished by mapping the desired perceived depth of an image component into spectral color. Depth decoding is performed by an optical system which shifts the spatial positions of the colors in the image to create left and right views. The process is particularly well suited to the creation of stereoscopic laser shows. Other applications are also being pursued.
Conference Paper
In this paper we present a user study on the use of Depth of Field for depth perception in Direct Volume Rendering. Direct Volume Rendering with Phong shading and perspective projection is used as the baseline. Depth of Field is then added to see its impact on the correct perception of ordinal depth. Accuracy and response time are used as the metrics to evaluate the usefulness of Depth of Field. The onsite user study has two parts: static and dynamic. Eye tracking is used to monitor the gaze of the subjects. From our results we see that though Depth of Field does not act as a proper depth cue in all conditions, it can be used to reinforce the perception of which feature is in front of the other. The best results (high accuracy & fast response time) for correct perception of ordinal depth occurs when the front feature (out of the two features users were to choose from) is in focus and perspective projection is used.
Article
Accurately and automatically conveying the structure of a volume model is a problem not fully solved by existing volume rendering approaches. Physics-based volume rendering approaches create images which may match the appearance of translucent materials in nature, but may not embody important structural details. Transfer function approaches allow flexible design of the volume appearance, but generally require substantial hand tuning for each new data set in order to be effective. We introduce the volume illustration approach, combining the familiarity of a physics-based illumination model with the ability to enhance important features using nonphotorealistic rendering techniques. Since features to be enhanced are defined on the basis of local volume characteristics rather than volume sample value, the application of volume illustration techniques requires less manual tuning than the design of a good transfer function. Volume illustration provides a flexible unified framework for enhancing structural perception of volume models through the amplification of features and the addition of illumination effects.
Article
The effective visualization of vascular structures is critical for diagnosis, surgical planning as well as treatment evaluation. In recent work, we have developed an algorithm for vessel detection that examines the intensity profile around each voxel in an angiographic image and determines the likelihood that any given voxel belongs to a vessel; we term this the "vesselness coefficient" of the voxel. Our results show that our algorithm works particularly well for visualizing branch points in vessels. Compared to standard Hessian based techniques, which are fine-tuned to identify long cylindrical structures, our technique identifies branches and connections with other vessels. Using our computed vesselness coefficient, we explore a set of techniques for visualizing vasculature. Visualizing vessels is particularly challenging because not only is their position in space important for clinicians but it is also important to be able to resolve their spatial relationship. We applied visualization techniques that provide shape cues as well as depth cues to allow the viewer to differentiate between vessels that are closer from those that are farther. We use our computed vesselness coefficient to effectively visualize vasculature in both clinical neurovascular x-ray computed tomography based angiography images, as well as images from three different animal studies. We conducted a formal user evaluation of our visualization techniques with the help of radiologists, surgeons, and other expert users. Results indicate that experts preferred distance color blending and tone shading for conveying depth over standard visualization techniques.
Article
The management of unruptured intracranial aneurysms is controversial. Investigators from the International Study of Unruptured Intracranial Aneurysms aimed to assess the natural history of unruptured intracranial aneurysms and to measure the risk associated with their repair. Centres in the USA, Canada, and Europe enrolled patients for prospective assessment of unruptured aneurysms. Investigators recorded the natural history in patients who did not have surgery, and assessed morbidity and mortality associated with repair of unruptured aneurysms by either open surgery or endovascular procedures. 4060 patients were assessed-1692 did not have aneurysmal repair, 1917 had open surgery, and 451 had endovascular procedures. 5-year cumulative rupture rates for patients who did not have a history of subarachnoid haemorrhage with aneurysms located in internal carotid artery, anterior communicating or anterior cerebral artery, or middle cerebral artery were 0%, 2. 6%, 14 5%, and 40% for aneurysms less than 7 mm, 7-12 mm, 13-24 mm, and 25 mm or greater, respectively, compared with rates of 2 5%, 14 5%, 18 4%, and 50%, respectively, for the same size categories involving posterior circulation and posterior communicating artery aneurysms. These rates were often equalled or exceeded by the risks associated with surgical or endovascular repair of comparable lesions. Patients' age was a strong predictor of surgical outcome, and the size and location of an aneurysm predict both surgical and endovascular outcomes. Many factors are involved in management of patients with unruptured intracranial aneurysms. Site, size, and group specific risks of the natural history should be compared with site, size, and age-specific risks of repair for each patient.