DISTANCE-ADAPTED 2D MANIPULATION TECHNIQUES

FOR LARGE HIGH-RESOLUTION DISPLAY ENVIRONMENTS

Anke Lehmann and Oliver Staadt

Institute for Computer Science, University of Rostock, Albert-Einstein-Str. 22, 18059 Rostock, Germany

Keywords:

Large High-resolution Display, 2D Manipulation, 3D User Interaction.

Abstract:

Physical navigation presents an intuitive overview+detail technique in large high-resolution display environ-

ments where users interact at different distances. Thus, the user interaction should be adapted to the user’s

current position. This ongoing work presents bimanual interaction techniques that use distance-adapted map-

ping of physical hands motion to virtual relative objects motion (named interaction scaling). Interaction scaling

allows the user to manipulate high-resolution content with the appropriate accuracy at diverse distances. In

our initial experiment, we considered two relative mapping functions: continuous mapping and zone mapping.

The mapping functions manipulate the 2D position of the virtual cursor. The parameters for the 2D manipu-

lation tasks (i.e., scaling, rotation, translation) are calculated from the virtual cursor behavior. We developed

two bimanual interaction techniques for 2D object manipulation that support the interaction scaling: separated-

cursors and connected-cursors method. The separated-cursors technique differentiates the users hands between

dominant hand and nondominant hand. The connected-cursors technique uses an additional midpoint cursor

that is located between both virtual cursors. In an explorative study we evaluated the separated-cursors tech-

nique with and without interaction scaling. An interesting result is that the participants often needed less object

selections to sort objects with interaction scaling compared to the scenario without interaction scaling.

1 INTRODUCTION

Large high-resolution displays (LHRD) combine a

large physical display area with high pixel density.

On the one hand LHRDs allow users to perceive

the global context of complex information by sim-

ply stepping away. On the other hand the users are

able to step closer to the display to see detailed in-

formation. In this context, interaction techniques for

LHRDs should support mobility, accuracy and easy

availability to carry out interaction tasks.

The interaction techniques described in this paper

are developed in the context of our Tele-Presence Sys-

tem comprising a 24-panel LCD wall (Willert et al.,

2010). The setting allows us to display remote users at

a natural scale, at a sufﬁcient resolution, and to share

interactive high-resolution content. For collaboration

work, a shared virtual space with variable depth is lo-

cated between the local and remote user.

For this environment we require intuitive user in-

teraction that supports the large interaction volume in

front of the LCD wall. The user should be able to

carry out distant selection and manipulation as well

as manipulation with high precision at close range.

In general the user’s distance in front of a large dis-

play is considered for selection and navigation tasks

in previous work (Kopper et al., 2010; Peck et al.,

2009; Vogel and Balakrishnan, 2005).

In our ongoing work we investigate how the user

beneﬁts from the user–display distance for simple 2D

manipulation tasks. The main contribution of our

work is the interaction scaling approach that allows

the user to manipulate 2D objects with the appropri-

ate accuracy in varying distance.

We developed two bimanual interaction tech-

niques for 2D object manipulation: separated-cursors

and connected-cursors method. In an explorative

study, we evaluated one distance-adapted 2D manip-

ulation technique (separated-cursors) with two map-

ping functions, which map the physical absolute

hands motion to virtual relative objects motion.

2 RELATED WORK

In large high-resolution display environments interac-

tion techniques have to be customized that the user is

able to interact with virtual content from various dis-

387

Lehmann A. and Staadt O..

DISTANCE-ADAPTED 2D MANIPULATION TECHNIQUES FOR LARGE HIGH-RESOLUTION DISPLAY ENVIRONMENTS.

DOI: 10.5220/0003857903870394

In Proceedings of the International Conference on Computer Graphics Theory and Applications (GRAPP-2012), pages 387-394

ISBN: 978-989-8565-02-0

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

tances.

For instance, extended laser pointer techniques al-

low the user to interact with (physically) unreachable

objects and small objects in large display environ-

ments (Peck, 2001; Ahlborn et al., 2005; Argelaguet

and Andujar, 2009; Fukazawa et al., 2010). Ahlborn

et al. (Ahlborn et al., 2005) have developed an effec-

tive detection algorithm of laser pointer interaction on

large-scale display environments that works at various

lighting conditions. Using a predictive searching re-

duces the amount of image processing to detect the

laser pointer dot. Using infrared laser pointer tech-

niques (Cheng and Pulo, 2003; K

onig et al., 2007)

the on-screen cursor is invisible, thus the problem

of hand jitter and latency are hidden from the user.

Those techniques use a visual representation of the

invisible laser dot (e.g., hotspot, visual cursor). Vo-

gel and Balakrishnan (Vogel and Balakrishnan, 2005)

investigated techniques for pointing and clicking us-

ing only the user’s hand. For example, they developed

a hybrid pointing technique that uses ray casting for

coarse pointing interaction and relative pointing for

more precise interaction.

Another challenge is the manipulation of virtual

objects on large workspaces. In (Collomb et al., 2005)

dragging techniques use an acceleration method to

overcome large distances. Grossmann et al. (Gross-

man et al., 2001) provide a two-handed interaction

technique for the creation of 3D models by drawing

2D proﬁle curves on large displays. In (Jeon et al.,

2010) camera-equipped mobile phones are utilized by

the users to interact with large displays. The virtual

objects have been manipulated based on the motion

gestures that are performed by the user with his mo-

bile phone. The CGLXTouch project (Ponto et al.,

2011) supports multitouch device interaction by mul-

tiple users. A low-resolution screen image of the dis-

play wall is displayed on the multitouch device cor-

responding to the available pixel resolution of the de-

vice. Thus, users can indirectly manipulate data on

the ultra-high-resolution display via multitouch ges-

tures on their devices concurrently.

Previous work (Jota et al., 2009; Kopper et al.,

2010; Peck et al., 2009; Vogel and Balakrishnan,

2005) has shown that the current distance of a user

to a large display affects the usability and the user’s

performance. For example, when stepping closer to

LHRDs it is helpful to reﬁne the selection radius or

to adjust the navigation speed. In (Kopper et al.,

2008) the user is able to switch between absolute and

relative mapping for pointing interaction. Further-

more, the user can control a zoom window. It pro-

vides a magniﬁed view of objects, where the user can

precisely interact. Multiscale interaction by Peck et

Figure 1: Interaction scaling with a) continuous mapping

(left) and b) zone mapping (right).

al. (Peck et al., 2009) is a 3D interaction technique

that uses the beneﬁt of physical navigation in front

of a large high-resolution display. Depending on the

user’s distance from the interaction object, he con-

trols the scale of interaction in addition to the scale

of perception. For example, when the user changes

his physical position relative to the display, the mul-

tiscale cursor is changed, which rescales the data on

the display interactively. The interactive public am-

bient display (Vogel and Balakrishnan, 2004) utilizes

the user distance to control public and personal infor-

mation presentation. In (Shoemaker et al., 2010) the

user distance controls the virtual light-source position

that affects the interactive user’s shadow cast.

Our objective is to ﬁnd a general approach for

scalable interaction that adjusts the interaction pre-

cision to the user’s position in front of a large high-

resolution display.

3 INTERACTION SCALING

Interaction scaling means to adjust the technical pre-

cision of user interaction depending on the user’s dis-

tance to the LHRD. Our objective is to ﬁnd an ad-

equate mapping of the user’s movements to virtual

movements. Interaction scaling allows the user to ma-

nipulate high-resolution content with the appropriate

accuracy in varying distance.

We propose a customized mapping of the gesture

to the virtual cursor movements taking into consid-

eration the user’s distance to the LHRD. The current

distance is calculated using the average of the distance

of both hands. For example, if the user stands farther

away from the display he should be able to move an

object across a large LCD wall with a small gesture.

In contrast, if he steps closer to the display, the virtual

object should be moved with high accuracy by using

large physical movements.

In our initial experiment, we consider two relative

mapping functions: continuous mapping and zone

mapping. The mapping functions are applied to the

GRAPP 2012 - International Conference on Computer Graphics Theory and Applications

388

Figure 2: Interaction techniques: a) separated-cursors (left), b) connected-cursors (right). The crosshair cursor indicates the

selected object.

virtual cursors as visual representation of the user’s

hand.

The continuous mapping function applies a linear

factor to the physical movements. The mapping fac-

tor is equal to one at middle distance of the interac-

tion space. The factor decreases linearly closer to the

LHRD and grows linearly with increasing distance

(see Figure 1, left). We use a simple linear equation

to calculate the mapping factor (m f ) depending on the

current user–display distance (udd); m and b are des-

ignate constants:

m f = f (udd) = m udd + b

For zone mapping we divide the interaction space

in front of the LHRD into four regions with prede-

ﬁned mapping factors: close-up range, mid-range,

distant and farther region. With zone mapping the

mapping factor is increased in each region away from

the display (see Figure 1, right). There is no distance

where the physical movements are equal to virtual

movements as opposed to continuous mapping. In the

ﬁrst two regions the mapping factor is less than one.

In these regions the user is able to manipulate objects

precisely. In the other two regions the user handles

object with coarse precision. Here, the mapping fac-

tor is greater than one.

Interaction scaling just manipulates the 2D posi-

tion of the virtual cursor. The distance-related map-

ping factor is multiplied by the absolute hand posi-

tion to determine the corresponding virtual cursor po-

sition. The parameters for the manipulation tasks (i.e.,

scaling, rotation, translation) are calculated by the vir-

tual cursors behavior. Thus, the mapping factor in-

ﬂuences the precision of the manipulation indirectly.

Similar to common two ﬁnger multitouch gestures,

the scaling factor is calculated based on the distance

change between both cursors. The rotation angle is

computed from the angular rotation of the line be-

tween both cursors.

We developed two bimanual interaction tech-

niques for 2D object manipulation that support the

interaction scaling: separated-cursors and connected-

cursors. We use tracked wireless input devices with

multiple buttons to simplify the triggering of user in-

teraction. Thus, the user is able to point at an object,

push the button to start the interaction, and release the

button to complete the interaction.

The separated-cursors technique differentiates the

user’s hands between dominant and nondominant

hand. The dominant hand is used to select and trans-

late an object. The nondominant hand is used to scale

and rotate the selected object (see Figure 3, left). Dur-

ing scale/rotate mode the dominant hand’s virtual cur-

sor is ﬁxed and the corresponding hand motion is ig-

nored. The object manipulation is illustrated in Fig-

ure 2 (left).

The connected-cursors technique uses an addi-

tional midpoint cursor that is located between both

virtual cursors (see Figure 2, right). This midpoint

cursor is used to select an object. In order to trans-

late a selected object the virtual cursors have to move

to the desired direction simultaneously. Increasing or

decreasing the distance between the cursors resizes a

virtual object. To rotate an object the user moves his

hands in a fashion similar to turning a steering wheel.

The connected-cursors method allows the user to se-

lect and manipulate a virtual object in a continuous

manner without changing the interaction mode. The

manipulation tasks are performed simultaneously.

4 EXPLORATORY STUDY

We conducted an experiment to evaluate our interac-

tion scaling techniques for LHRDs. Our hypothesis

is that interaction scaling improves the user perfor-

mance compared to manipulation tasks that do not

exploit the current user distance to the large display

wall.

In an explorative study, we investigate one interac-

tion technique (separated-cursors) with two mapping

functions of interaction scaling (continuous mapping,

DISTANCE-ADAPTED 2D MANIPULATION TECHNIQUES FOR LARGE HIGH-RESOLUTION DISPLAY

ENVIRONMENTS

389

Figure 3: Application with separated-cursors technique: the differentiation of user’s hands (left) and the target box of the

experiment (right).

zone mapping), as described in Section 3. The pri-

mary function of this study is to discover parameters,

which affect the application of interaction scaling for

2D manipulation tasks. Using these results we are

able to develop a suitable interaction scaling depend-

ing on the LHRD setup and manipulation task in order

to improve the user performance.

In the study, we would like to evaluate just one

interaction technique to reduce the duration of an ex-

periment for each participant. After initial tests, we

expect that the separated-cursors technique provides

promising results.

We implemented a 2D puzzle solver application to

compare the task completion time and the number of

object selection with and without interaction scaling.

We use OpenGL for rendering and the Vrui toolkit for

device and display handling

4.1 Apparatus

We use a ﬂat tiled LCD wall consisting of 24 DELL

2709W displays, where each tile has a resolution of

1920 x 1200 pixels, with a total resolution of 11520

x 4800 (55 million pixels). In front of the display

wall the interaction space is approximately 3.7 x 2.0

x 3.5 meter. The user holds a Nintendo Wii Remote

in each hand to interact with the application. We use

an infrared tracking system (Naturalpoint OptiTrack

TrackingTools) with 12 cameras arranged semicircu-

larly around the LCD wall. The current user–display

distance is calculated by the average of the distance

of both tracked input devices.

Pointing gesture allows an intuitive interaction

with virtual objects at different distances. In order to

simplify future tracking the hands position are tracked

http://idav.ucdavis.edu/∼okreylos/ResDev/Vrui/index.

html

only in front of the LHRD and we do not consider the

orientation of the hands or input devices. Therefore,

3D position detection is quite possible by various op-

tical tracking methods. The hands’ positions are pro-

jected orthogonally on the screen.

After a while a drift arises between absolute hand

motion and relative cursor movements by using inter-

action scaling. Since the mapping factor is multiplied

by the absolute hand position to determine the virtual

relative cursor position. For this reason, the user can

trigger a reset of the virtual cursors position corre-

sponding to the hands position.

Since we use an LCD wall with bezels, the puz-

zle application is conﬁgured in such a way that no

virtual content is occluded by the bezels. The virtual

objects are limited by a minimum and maximum size

(range size 1cm–50cm). The size limits were deter-

mined experimentally and the virtual objects are visi-

ble at distance. The virtual objects are generated with

random size and position at application start without

overlapping. The target objects will always be in the

same place on the display and be of the same size.

The tracking system setup restricts the user interac-

tion distance from 40 centimeters to 3 meters.

There are two interaction scaling scenarios: con-

tinuous mapping and zone mapping. Continuous

mapping calculates a mapping factor that increases

from 0.2 to 2.5 with increasing distance. The zone

mapping scenario divides the interaction space into

four even regions with a depth of 50 centimeters.

The experimental mapping factors are 0.2 (close-up

range), 0.7 (mid-range), 1.5 (distant region), and 2.5

(farther region). If the user moves his hand ten cen-

timeters in physical space at close-up range then the

corresponding cursor is moved three centimeters in

virtual space. Conversely, the virtual cursor is moved

25 centimeters while the physical hand movement is

GRAPP 2012 - International Conference on Computer Graphics Theory and Applications

390

ten centimeters at farther region. For the purpose of

comparison the scenario without interaction scaling

maps the real hands movement directly to the virtual

cursors motion.

4.2 Procedure and Design

The experiment consists of two parts - a training

phase and the test scenarios. At the beginning the

subject performs a 10 minutes tutorial to practice the

interaction tasks with and without interaction scaling.

Afterwards the evaluation starts. The task is to sort

eight objects (a couple of triangles, circles, squares,

and stars) in the corresponding targets (see Figure 3,

right). In the tutorial the subject sorts three various

object types. The objects position, size and orienta-

tion are randomized. The objects have to be manipu-

lated to ﬁt into the targets by dragging, scaling and ro-

tating. Only when the user drops down the object onto

the desired target by releasing the selection button, the

application will test the ﬁtting accuracy. The object

tolerance is two millimeters regarding target size, po-

sition and orientation. Object orientation is described

by the position of the vertices. The tolerance value

was determined experimentally in order that users are

able to sort objects without frustrating. If an object is

sorted correctly, it will be locked to avoid undesired

movements.

The trial starts with the ﬁrst object selection, and

ﬁnishes with the last sorted object. The experimental

software measures the task completion time in sec-

onds and the number of object selections. After the

test, the subject ﬁlls in a subjective questionnaire.

We used a within-subject design. The independent

variable was the interaction scaling with three levels

(without mapping, continuous mapping, zone map-

ping). The presentation order was counter-balanced.

Each participant performed three scenarios with two

repetitions. The dependent variable was the task com-

pletion time and the number of object selections. The

subject was informed whether interaction scaling is

active or not.

The explorative study includes 15 participants but

the test results of two persons were incomplete. Thus,

we can only evaluate the results of 13 subjects. All

13 participants were volunteers that were college stu-

dents (54%) or staff members (46%) from computer

science department. The participants were 31% fe-

male and 69% male. The ages of the participants

ranged from 20 years old to 39 years old with an av-

erage age of 27. The participants had less or no prior

experience working with large display environments.

Figure 4: Boxplot of completion times within the scenarios

in the ﬁrst iteration.

4.3 Results

We performed a one-way ANOVA to determine sta-

tistical signiﬁcance in the preliminary study. The pa-

rameter completion time shows no interesting results

or no statistically signiﬁcant differences between sce-

narios. In general, the completion time is larger at

the ﬁrst try to solve the puzzle. About ten partici-

pants (77%) improved their task completion time in

the second run of the scenario. In the scenario with-

out interaction scaling the mean improvement was 47

seconds and in the scenarios with interaction scaling

on average 20 seconds. We observed that the presen-

tation order of the scenarios has an inﬂuence on the

completion time. The puzzle generates a low training

curve. The more the participant practiced the puzzle

the faster he was able to solve the puzzle without in-

teraction scaling.

There is a correlation between task completion

time and number of used object selections. An in-

crease in completion time is combined with a larger

number of selections. Sometimes the user needed

more time to sort in a speciﬁc object. For instance,

the subject required 14 object selections to sort in one

object and then only two or ﬁve selections to sort an-

other object in the same scenario. We observed that

the completion time is a less important parameter of

user performance in our experiment.

In general, the participants were faster on average

without interaction scaling (mean: 209 seconds, mean

deviation: 68 seconds) in comparison to the continu-

ous mapping scenario (mean: 223 seconds, mean de-

viation: 85 seconds) and the zone mapping scenario

(mean: 268 seconds, mean deviation: 116 seconds).

The measured completion times are illustrated in Fig-

ure 4.

We identify interesting results in terms of used ob-

ject selections. In particular, a difference is recog-

DISTANCE-ADAPTED 2D MANIPULATION TECHNIQUES FOR LARGE HIGH-RESOLUTION DISPLAY

ENVIRONMENTS

391

Table 1: Frequency of used object selections in each scenario in the ﬁrst run.

scenario without mapping continuous mapping zone mapping

selections frequency of object selections (absolute and relative values)

1 10 9,6% 11 10,6% 11 10,6%

2 27 26,0% 34 32,7% 31 29,85%

3 16 15,4% 17 16,3% 13 12,5%

4 15 14,4% 5 4,8% 9 8,6%

5 9 8,6% 7 6,7% 9 8,6%

> 5 27 26,0% 30 28,9% 31 29,85%

Table 2: Frequency of used object selections for each object type and scenario (without mapping (wm), continuous mapping

(cm), zone mapping (zm)) in the ﬁrst run.

object type squares circles triangles stars

selections wm cm zm wm cm zm wm cm zm wm cm zm

1 2 1 4 3 3 1 1 0 2 4 7 4

2 6 9 6 10 8 10 7 11 9 4 6 6

3 4 7 2 4 5 6 5 2 1 3 3 4

4 6 0 4 3 1 1 3 2 1 3 2 3

> 4 8 9 10 6 9 8 10 11 13 12 8 9

nizable between using interaction scaling and without

interaction scaling in the ﬁrst iteration. The number

of used object selections in each scenario is shown in

Figure 5. We captured 104 samples (8 objects x 13

subjects) within one run. Table 1 shows that more ob-

jects were sorted with two selections with interaction

scaling by comparison without interaction scaling.

To select an object the user has to move the cor-

responding virtual cursor of his dominant hand onto

the object and press the Wii Remote button. The

user moves a selected object by moving his dominant

hand. If the user presses the Wii Remote button of the

nondominant hand then the scale/rotate mode is ac-

tivated until the button is released. This mode allows

the user to scale and rotate the selected object by mov-

ing his nondominant hand. The object is selected until

the button of the dominant hand is released. However,

a statistical signiﬁcance between the scenarios is not

present in the preliminary study.

In addition, Table 2 indicates the difference be-

tween the object types. For instance, rotatable objects

(e.g., square, triangle, star) were sorted with interac-

tion scaling (continuous mapping and zone mapping)

more frequently compared to the scenario without

mapping. We observed that most of subjects moved

an object closer to the corresponding target at distant

and then they stepped closer to the display to adjust

the object. Using this strategy, the subjects needed

only two selections within interaction scaling scenar-

ios. We observed that when participants used more

than ﬁve object selections they were impatient or they

did not go closer to the display.

The repetitions of the scenario without interaction

scaling more participants were also able to sort the

objects with two selections. This suggests that ob-

ject sorting without interaction scaling is a matter of

practice. However, a lot of exercise is not required be-

cause the learning curve is low and ﬁnished after three

solved puzzles. This observation is also reﬂected by

the subject comments.

Figure 5: Boxplot of used object selections within the sce-

narios in the ﬁrst iteration.

In the zone mapping scenario many users had dif-

ﬁculties solving the puzzle. The precision of the ma-

nipulation was very high in the close-up range; the

participant took more physical movements to manip-

ulate the objects. In addition, in the zone mapping

scenario the precision factor changed rapidly because

of the small regions. In the continuous mapping sce-

nario the ﬂuent adjustment was less perceived by the

users.

Furthermore, we observed that the drift between

hands and cursors unfavorably affects the partici-

pants. In the zone mapping scenario, the difference

GRAPP 2012 - International Conference on Computer Graphics Theory and Applications

392

between absolute hand position and relative cursor

position is perceptible by the user, due to the rapid

changes of the mapping factor. In the continuous

mapping scenario the mapping factor changes con-

tinuously. Here, the drift is less pronounced and the

users are able to compensate the effect. Users with

virtual reality experience were able to compensate the

drift and they speciﬁcally applied the cursor reset,

whereas users without experience in virtual reality en-

vironments were dissatisﬁed with the drift effect.

Experimental tests with a few users indicated that

the completion time improves considerably by exer-

cise with interaction scaling. Primarily, users with

a natural hand tremor could beneﬁt from interaction

scaling. This idea is supported by the participants’

comments. For example, eight participants preferred

the scenario with interaction scaling. Some users de-

sired to interactively disable the automatic adjustment

of the precision. For instance, then the user is able to

resize quickly large objects to ﬁt into a small target at

close-up range.

5 CONCLUSIONS AND FUTURE

WORK

We presented distance-adapted 2D manipulation tech-

niques that adjust the interaction precision to the

user’s position in front of a large high-resolution dis-

play. This interaction scaling allows the user to per-

form distant selection and manipulation in varying

distance with the appropriate accuracy.

The introduced explorative pilot study is supposed

to prepare the user study, particularly by develop-

ing and attempting the suitable interaction scaling

for 2D manipulation tasks. Therefore, we evaluated

the separated-cursors technique with two mapping

functions of interaction scaling (continuous mapping,

zone mapping) and without interaction scaling. We

implemented a 2D puzzle solver application to com-

pare the task completion time and number of used ob-

ject selections.

The preliminary study shows that the users accept

interaction scaling. In particular, they preferred the

higher precision close to the display and the reducing

of their hand tremor. The results demonstrate that the

users were on average faster with continuous mapping

compared to zone mapping. In addition, many par-

ticipants completed the task faster without interaction

scaling. One possible explanation is the short training

curve of the puzzle application. The more the subjects

practiced the puzzle the faster they were able to solve

it.

An interesting result is that the participants needed

less object selections to sort objects with interaction

scaling compared to the scenario without interaction

scaling. In particular, the subjects were more often

able to sort objects at most two selections with in-

teraction scaling scenarios. However, the number of

used selections has shown that users required less ob-

ject selections but generally they took more time to

solve the puzzle with interaction scaling. Here, it

needs to be checked, if more complex tasks or other

tasks will improve the task completion time by using

interaction scaling.

The explorative study indicates that the cursor

drift strongly affects the user’s performance. In par-

ticular, users without virtual reality experience were

dissatisﬁed with the drift in the zone mapping sce-

nario. The drift effect is less pronounced in the con-

tinuous mapping scenario than in the zone mapping

scenario, because the mapping factor changes contin-

uously and the user is able to compensate the differ-

ence between hands motion and cursors motion.

The reason we did not get statistical signiﬁcant re-

sults, we suppose that the drift negatively affects the

results. Thus, the drift has abolished the beneﬁts of

interaction scaling.

As future work we plan to involve intention recog-

nition to obviate the drift between absolute hand mo-

tion and relative cursor motion. So the interaction pre-

cision is adjusted to the user’s current position and his

motion sequence. The drift problem occurs at both

interaction techniques (separated-cursors, connected-

cursors) and the available cursor reset is less intuitive.

Thus, we will evaluate the connected-cursors tech-

nique with the reduced drift effect.

In our LHRD environment a maximum mapping

factor of one is sufﬁciently good to reach objects from

a distance. We expect using deeper zones; the differ-

ences between the mapping methods are clearly visi-

ble. So interaction scaling with zone mapping could

improve the user’s performance. Additionally we will

readjust the tracking volume that the user is able to

interact closer at the display. We will also test to cal-

culate user–display distance based on the head posi-

tion to adjust the precision factor as expected from

the user, instead of using the average of the distance

of both hands.

The next steps include utilizing the results of the

explorative study (i.e., to minimize the drift, complex

task, appropriate interaction zones, etc.). We will per-

form a user study to evaluate how the user beneﬁts

from the user–display distance for 2D manipulation

tasks.

DISTANCE-ADAPTED 2D MANIPULATION TECHNIQUES FOR LARGE HIGH-RESOLUTION DISPLAY

ENVIRONMENTS

393

ACKNOWLEDGEMENTS

This work was supported by EFRE fond of the Eu-

ropean Community and a grant of the German Na-

tional Research Foundation (DFG), Graduate School

1424 MuSAMA. We thank the anonymous reviewers

for their valuable contribution.

REFERENCES

Ahlborn, B. A., Thompson, D., Kreylos, O., Hamann, B.,

and Staadt, O. G. (2005). A practical system for laser

pointer interaction on large displays. In Proceedings

of the ACM symposium on Virtual reality software

and technology, pages 106–109, Monterey, CA, USA.

ACM.

Argelaguet, F. and Andujar, C. (2009). Efﬁcient 3d pointing

selection in cluttered virtual environments. Computer

Graphics and Applications, IEEE, 29(6):34 –43.

Cheng, K. and Pulo, K. (2003). Direct interaction with

large-scale display systems using infrared laser track-

ing devices. In Proceedings of the Asia-Paciﬁc sympo-

sium on Information visualisation - Volume 24, APVis

’03, pages 67–74, Darlinghurst, Australia, Australia.

Australian Computer Society, Inc.

Collomb, M., Hasco

et, M., Baudisch, P., and Lee, B.

(2005). Improving drag-and-drop on wall-size dis-

plays. In Proceedings of Graphics Interface 2005, GI

’05, pages 25–32, School of Computer Science, Uni-

versity of Waterloo, Waterloo, Ontario, Canada. Cana-

dian Human-Computer Communications Society.

Fukazawa, R., Takashima, K., Shoemaker, G., Kitamura, Y.,

Itoh, Y., and Kishino, F. (2010). Comparison of mul-

timodal interactions in perspective-corrected multi-

display environment. In 3D User Interfaces (3DUI),

2010 IEEE Symposium on, pages 103 –110.

Grossman, T., Balakrishnan, R., Kurtenbach, G., Fitzmau-

rice, G., Khan, A., and Buxton, B. (2001). Interaction

techniques for 3d modeling on large displays. In ACM

I3DG 1999 Symposium on Interactive 3D Graphics,

pages 17–23.

Jeon, S., Hwang, J., Kim, G. J., and Billinghurst, M.

(2010). Interaction with large ubiquitous displays us-

ing camera-equipped mobile phones. Personal Ubiq-

uitous Computing, 14(2):83–94.

Jota, R., Pereira, J. a. M., and Jorge, J. A. (2009). A com-

parative study of interaction metaphors for large-scale

displays. In Proceedings of the 27th international

conference extended abstracts on Human factors in

computing systems, CHI ’09, pages 4135–4140, New

York, NY, USA. ACM.

onig, W. A., Bieg, H.-J., Schmidt, T., and Reiterer, H.

(2007). Position-independent interaction for large

high-resolution displays. In IHCI’07: Proceedings

of IADIS International Conference on Interfaces and

Human Computer Interaction 2007, pages 117–125.

IADIS Press.

Kopper, R., Bowman, D. A., Silva, M. G., and McMahan,

R. P. (2010). A human motor behavior model for dis-

tal pointing tasks. International Journal of Human-

Computer Studies, 68(10):603 – 615.

Kopper, R., Silva, M. G., McMahan, R. P., and Bowman,

D. A. (2008). Increasing the Precision of Distant

Pointing for Large High-Resolution Displays. Tech-

nical report, Computer Science, Virginia Tech.

Peck, C. H. (2001). Useful parameters for the design of

laser pointer interaction techniques. In CHI ’01 ex-

tended abstracts on Human factors in computing sys-

tems, CHI ’01, pages 461–462, New York, NY, USA.

ACM.

Peck, S. M., North, C., and Bowman, D. (2009). A mul-

tiscale interaction technique for large, high-resolution

displays. In Proceedings of the 2009 IEEE Sympo-

sium on 3D User Interfaces, 3DUI ’09, pages 31–38,

Washington, DC, USA. IEEE Computer Society.

Ponto, K., Doerr, K., Wypych, T., Kooker, J., and

Kuester, F. (2011). Cglxtouch: A multi-user multi-

touch approach for ultra-high-resolution collaborative

workspaces. Future Generation Computer Systems,

27(6):649–656.

Shoemaker, G., Tsukitani, T., Kitamura, Y., and Booth,

K. S. (2010). Body-centric interaction techniques

for very large wall displays. In Proceedings of the

6th Nordic Conference on Human-Computer Inter-

action: Extending Boundaries, NordiCHI ’10, pages

463–472, New York, NY, USA. ACM.

Vogel, D. and Balakrishnan, R. (2004). Interactive pub-

lic ambient displays: transitioning from implicit to

explicit, public to personal, interaction with multiple

users. In Proceedings of the 17th annual ACM sympo-

sium on User interface software and technology, UIST

’04, pages 137–146, New York, NY, USA. ACM.

Vogel, D. and Balakrishnan, R. (2005). Distant freehand

pointing and clicking on very large, high resolution

displays. In Proceedings of the 18th annual ACM

symposium on User interface software and technol-

ogy, UIST ’05, pages 33–42, New York, NY, USA.

ACM.

Willert, M., Ohl, S., Lehmann, A., and Staadt, O. (2010).

The Extended Window Metaphor for Large High-

Resolution Displays. In JVRC10: Joint Virtual Reality

Conference of EGVE - EuroVR - VEC, pages 69–76,

Stuttgart, Germany. Eurographics Association.

GRAPP 2012 - International Conference on Computer Graphics Theory and Applications

394