scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Passive yet Expressive TouchTokens

TL;DR: A new recognizer, that analyzes the micro-movements of the fingers that hold the tokens, enables the system to detect when a token is left on the surface rather than taken off it, and can also detect bend events that can be mapped to command triggers, and a squeezed state that could be used for quasi-modal interaction.
Abstract: TouchTokens are passive tokens that can be recognized on any capacitive surface based on the spatial configuration of the fingers that hold them. However, interaction with these tokens is confined to the basic two-state model of touch interaction as the system only knows the tokens' position and cannot detect tokens that are not touched. We increase the expressive power of TouchTokens by introducing laser-cut lattice hinges in their design, so as to make them flexible. A new recognizer, that analyzes the micro-movements of the fingers that hold the tokens, enables the system to detect when a token is left on the surface rather than taken off it. It can also detect bend events that can be mapped to command triggers, and a squeezed state that can be used for quasi-modal interaction.

Summary (3 min read)

INTRODUCTION

  • TouchTokens [9] provide a simple means to develop tangible interfaces.
  • The approach relies on easy-to-make passive tokens that feature notches constraining how users grasp them.
  • The definitive version will be published in CHI ’17, Denver, CO, USA.
  • The tokens flexible by introducing lattice-hinges in their design, and on the software side on a novel recognizer that analyzes the micro-movements of the token-holding fingers that remain in contact with the surface.
  • Finally, the authors describe their recognizer, and evaluate its performance.

MAKING TOUCHTOKENS MORE EXPRESSIVE

  • The authors achieve this with a novel design that makes the tokens flexible, and with an analysis of the micro-movements users make when performing these interactions, following an approach similar to the recognizers designed to detect thumb-tip micro-gestures [2, 10].
  • This section describes their new tokens and introduces their hypotheses regarding the micro-movements the authors expect to observe.

Designing Flexible TouchTokens

  • Figure 1 shows their novel set of tokens, which can be squeezed or bent by pinching them.
  • In their case, the authors performed several design iterations so as to make the tokens comfortable to manipulate while ensuring enough robustness.
  • The kerfs’ width, length and interspacing provide enough elasticity to make the tokens easy to deform without requiring too high a force, while ensuring that they revert to their original shape when not pinched.
  • The system does not even know whether a token has been left on the surface or removed off it.
  • In the latter case, the authors should observe finger traces that either remain still or move slightly toward the touch points’ centroid.

Squeezing Tokens

  • When squeezing a token, the user’s fingers remain in contact with the surface throughout the corresponding micromovements.
  • The authors hypothesized that when squeezing, they would observe touch traces that move toward the touch points’ centroid, and away from it when un-squeezing.
  • These can be used respectively to trigger discrete events, and to enter quasi-modes.

Bending Tokens

  • Bending a token leads to a state where users are keeping only one finger in contact with the surface .
  • As all other token manipulations involve at least two fingers, the number of fingers could be a discriminating factor.
  • It is too permissive, as it may also match cases where users lift two fingers off, but leave the token flat on the surface .
  • Again, micro-movements may help us detect actual bending actions.
  • The authors should thus observe still traces before lift-off when bending, as opposed to traces that slightly move away from the centroid in the other case.

COLLECTING TOUCH TRACES

  • The authors collected multi-touch traces of users performing the three types of manipulations described above.
  • The authors goal was to gather data about the different finger micro-movements, and to identify criteria that could enable us to recognize the corresponding manipulation events.
  • The authors were particularly interested in the typical profile of point-to-centroid average distance time-series associated with these movements.

Procedure

  • All participants performed the 3 manipulation events: Click and Drag & Drop, Leave on vs. Lift off and Bend vs. Leave flat.
  • Presentation order was counterbalanced using a Latin Square.
  • The authors collected data involving sliding movements in 4 main DIRECTIONs: up, down, left, right.
  • In the first case, they had to lift their fingers off the surface but leave the token on it.
  • For each event type, trials are first blocked by ACTION, then by DIRECTION within each ACTION (Event1 and Event2), and by FINGERCOUNT within each DIRECTION block (Event2).

RECOGNIZERS

  • The authors main hypothesis was that the micro-movements of interest to us could be observed by looking at the fingers’ traces, that should move slightly toward, or away from, the token’s center.
  • Parameter values (in bold) are determined in the next section.
  • |B|}, dre f −di > dsqz where dre f is the average distance in millimeters of a touch point to the centroid of the corresponding multi-touch sample when users register the token, and B is a buffer containing the successive values of d over the last buffersqz milliseconds.
  • This entails that their recognizer considers bent as a discrete event, that gets triggered only once users have unbent the token.
  • The criterion for bend is only evaluated after a time span of 100ms during which there has been exactly 1 contact point.

RECOGNIZER PARAMETERIZATION

  • For each of their three micro-movements, the authors measure the accuracy of their recognizer by running it on data collected for this micro-movement only.
  • The authors use the leave-one-out cross-validation technique to parameterize the recognizers: for each participant, they set the parameters to values that maximize the overall recognition score for the 11 other participants.
  • 2As a side note, the authors observed a recognition accuracy close to 90% for on/off states during informal tests using rigid tokens, suggesting that these micro-movements can also be detected on regular TouchTokens.

CONCLUSION

  • The authors new events enable developing more powerful interfaces where tokens can be dragged or clicked (bent, d), and where several tokens can be laid on the surface (on/off enabling the system to keep track of them).
  • This extended vocabulary can be used for different purposes, such as concurrently activating several filters, invoking commands on specific items or transferring data using drag-and-drop, click actions or contextual controls that take the tokens’ relative layout into account.

Did you find this useful? Give us your feedback

Content maybe subject to copyright    Report

HAL Id: hal-01562021
https://hal.inria.fr/hal-01562021
Submitted on 13 Jul 2017
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Passive yet Expressive TouchTokens
Rafael Morales González, Caroline Appert, Gilles Bailly, Emmanuel Pietriga
To cite this version:
Rafael Morales González, Caroline Appert, Gilles Bailly, Emmanuel Pietriga. Passive yet Expressive
TouchTokens. Proceedings of the 35th SIGCHI conference on Human Factors in computing systems,
May 2017, Denver, United States. pp.3741 - 3745, �10.1145/3025453.3025894�. �hal-01562021�

Passive yet Expressive TouchTokens
Rafael Morales González
1
Caroline Appert
1
Gilles Bailly
2,3
Emmanuel Pietriga
1
1
LRI, Univ. Paris-Sud, CNRS,
2
LTCI, CNRS,
3
Sorbonne Universités
INRIA, Telecom ParisTech, UPMC Univ Paris 06,
Université Paris-Saclay, Orsay, France Université Paris-Saclay, Paris, France CNRS, ISIR, Paris, France
ABSTRACT
TouchTokens are passive tokens that can be recognized on any
capacitive surface based on the spatial configuration of the fin-
gers that hold them. However, interaction with these tokens
is confined to the basic two-state model of touch interaction
as the system only knows the tokens’ position and cannot de-
tect tokens that are not touched. We increase the expressive
power of TouchTokens by introducing laser-cut lattice hinges
in their design, so as to make them flexible. A new recog-
nizer, that analyzes the micro-movements of the fingers that
hold the tokens, enables the system to detect when a token is
left on the surface rather than taken off it. It can also detect
bend events that can be mapped to command triggers, and a
squeezed state that can be used for quasi-modal interaction.
ACM Classification Keywords
H.5.2 : User Interfaces - Input devices and strategies.
Author Keywords
Tangible interaction; Multi-Touch input; Micro-movements
INTRODUCTION
TouchTokens [9] provide a simple means to develop tangible
interfaces. The approach relies on easy-to-make passive to-
kens that feature notches constraining how users grasp them.
Manipulating the tokens while maintaining the fingers in con-
tact with the touch-sensitive surface leads to specific multi-
touch spatial patterns that can be uniquely identified using a
relatively simple software recognizer. However, users are lim-
ited in how they can manipulate these tokens, as is often the
case with approaches based on capacitive sensing.
In this article, we aim at increasing the expressive power of
TouchTokens by making the system able to detect: 1) when a
token is left on or lifted off the surface, 2) when it is squeezed
and 3) when it is bent. We achieve this without introducing
any kind of instrumentation, thus preserving the simplicity
of the original approach, which relies exclusively on passive
tokens, and which works with any off-the-shelf capacitive sur-
face. Our solution relies on the hardware side on making
Rafael Morales González, Caroline Appert, Gilles Bailly & Emmanuel
Pietriga. Passive yet Expressive TouchTokens. In CHI ’17: Proceedings
of the 35th Annual ACM Conference on Human Factors in Computing
Systems, 3741-3745, ACM, May 2017.
©ACM, 2017. This is the authors version of the work. It is posted
here by permission of ACM for your personal use. Not for redistribu-
tion. The definitive version will be published in CHI ’17, Denver, CO, USA.
http://dx.doi.org/10.1145/3025453.3025894
the tokens flexible by introducing lattice-hinges in their de-
sign, and on the software side on a novel recognizer that ana-
lyzes the micro-movements of the token-holding fingers that
remain in contact with the surface.
After a short overview of related work, we describe the design
of our flexible tokens, based on lattice hinges which can easily
be obtained using fabrication processes such as laser cutting.
We then report on a formative study in which we collected a
sample of finger micro-movements that are representative of
the manipulations afforded by our flexible tokens. Finally, we
describe our recognizer, and evaluate its performance.
RELATED WORK
The most common approach to enabling tangible interaction
on surfaces that use diffuse illumination technology consists
in augmenting the objects with fiducial markers, and using
a vision-based algorithm to identify them and track their lo-
cation (see, e.g., [5]). Other projects have investigated tan-
gibles that reflect incoming light to the surface in a specific
way in order to support more manipulations, such as TZee
tangibles [14], which have the shape of a truncated pyramid
and support gesturing on their sides, or Lumino blocks [1],
which can be stacked. Diffuse illumination is a solution that
is usually reserved to large setups such as tabletops.
Another approach involves augmenting tangibles with mag-
nets. When coupled with a force-resistive screen, the system
can detect pressure and gestures performed on top of the to-
kens [6]. When coupled with a surface augmented with a
Hall sensor grid, the system can track tokens hovering over
the surface [8]. GaussBricks [7], which also rely on a display
equipped with Hall sensors, are bricks that can be assembled
together to create larger objects featuring both deformable
and rigid parts. While this approach enables very rich interac-
tions, it requires augmenting the surface with specific sensors,
and ensuring that the device’s environment is free of any fer-
rous object that could interfere with the tangibles’ magnetic
field.
Solutions based on capacitive sensing are more affordable,
but usually more limited. The system will often only be able
to track the tokens that users are touching. There are, how-
ever, a few exceptions that go beyond these limitations. Cap-
Stones and ZebraWidgets [3] are capacitive units that can be
assembled to configure different conductive circuits, enabling
more manipulations with the tangibles that can, for example,
be stacked or feature moving parts. PUCs [13] widgets rely
on the principle of mutual capacitance so as to be detected

(a) (c)(b)
Figure 1. Making a TouchToken flexible: (a) original, rigid TouchTo-
ken (circle, 4cm in diameter), (b) schematics of lattice-hinges, (c) flexible
TouchToken. Vector descriptions of all flexible TouchTokens available at
https://www.lri.fr/~appert/touchtokens/index.html.
even when users do not touch them. However, after a moment,
PUCs get rejected by the adaptive filtering method of capac-
itive screens. To avoid this issue, PERCs [12] are equipped
with sensors to capture the electrical field emitted by the ca-
pacitive screen, enabling them to know if they are on the
surface or not, and communicate their state (on vs. off the
surface) to the system via the Bluetooth protocol. Our con-
tribution also aims at increasing the number of possible inter-
actions with tokens but, as described in the next section, we
do so without relying on any advanced design or embedded
electronics.
MAKING TOUCHTOKENS MORE EXPRESSIVE
We contribute three novel primitives to the interaction vo-
cabulary of TouchTokens: a state (on/off ), a quasi-mode
(squeezed) and a discrete event (bent). We achieve this with
a novel design that makes the tokens flexible, and with an
analysis of the micro-movements users make when perform-
ing these interactions, following an approach similar to the
recognizers designed to detect thumb-tip micro-gestures [2,
10]. This section describes our new tokens and introduces
our hypotheses regarding the micro-movements we expect to
observe.
Designing Flexible TouchTokens
Figure 1 shows our novel set of tokens, which can be
squeezed or bent by pinching them. Laser-cutting lattice
hinges is a common method in the maker community to make
a piece of wood flexible using laser cutting. In our case, we
performed several design iterations so as to make the tokens
comfortable to manipulate while ensuring enough robustness.
The kerfs’ orientation was chosen so as to match that of the
comfortable pinch formed by the thumb on one side and the
{index, middle} couple of fingers on the other side. The kerfs’
width, length and interspacing provide enough elasticity to
make the tokens easy to deform without requiring too high a
force, while ensuring that they revert to their original shape
when not pinched. We also considered resistance to avoid ac-
cidental pinches during regular manipulations, and robustness
to avoid the risk of breaking.
Detecting Tokens’ on/off State
Making the system aware of whether a token is still on the
surface, or if it has been lifted off it, is an important feature of
tangible interaction. It allows users to lay out several tokens
on the surface (as in, e.g., Facet-streams [4]). Conductive
tokens usually rely on the fact that the human body is a con-
ductor. They thus become invisible to the system as soon as
(a) LEAVING ON
(b) LIFTING OFF
Figure 2. Finger micro-movements when leaving a token on the surface
(a), and when lifting it off (b).
(b) LEAVING FLAT
(a) BENDING
Figure 3. Micro-movements when (a) bending a token, (b) leaving it flat.
users no longer touch them. The system does not even know
whether a token has been left on the surface or removed off
it.
TouchTokens require users to both hold them by putting their
fingers in the notches and touch the surface with those fingers.
We hypothesized that the micro-movements made by the fin-
gers at the time they leave the surface would have a distinct
signature, depending on whether users were leaving tokens
on the surface or were lifting them off. Figure 2 illustrates
our hypothesis: when leaving a token on the surface, users
are likely going to relax their grasp, while when lifting it off,
they will likely maintain a firm grip, potentially compressing
the token a bit. In the former case, we should observe finger
traces that move slightly away from the touch points’ cen-
troid. In the latter case, we should observe finger traces that
either remain still or move slightly toward the touch points’
centroid.
Squeezing Tokens
When squeezing a token, the user’s fingers remain in con-
tact with the surface throughout the corresponding micro-
movements. We hypothesized that when squeezing, we
would observe touch traces that move toward the touch points’
centroid, and away from it when un-squeezing. If successful,
tokens can then be made to behave like a mouse with a but-
ton: quickly squeezing and releasing is equivalent to a click;
keeping the token squeezed and moving it on the surface is
equivalent to a drag. These can be used respectively to trig-
ger discrete events, and to enter quasi-modes.

Bending Tokens
Bending a token leads to a state where users are keeping only
one finger in contact with the surface (Figure 3-a). As all
other token manipulations involve at least two fingers, the
number of fingers could be a discriminating factor. However,
it is too permissive, as it may also match cases where users
lift two fingers off, but leave the token flat on the surface (Fig-
ure 3-b). Again, micro-movements may help us detect actual
bending actions. We hypothesize that users are likely going
to keep their index and middle fingers in contact with the to-
ken’s side when bending it, while they are going to relax their
grip when leaving it flat. We should thus observe still traces
before lift-off when bending, as opposed to traces that slightly
move away from the centroid in the other case.
COLLECTING TOUCH TRACES
We collected multi-touch traces of users performing the three
types of manipulations described above. Our goal was to
gather data about the different finger micro-movements, and
to identify criteria that could enable us to recognize the cor-
responding manipulation events. We were particularly inter-
ested in the typical profile of point-to-centroid average dis-
tance time-series associated with these movements.
Participants & Apparatus
Twelve volunteers (2 female), 23 to 40 year-old (avg. 28.83,
med. 28), participated in the data collection. They were
seated at a desk, manipulating tokens on a tablet (Samsung
SM-T810 Galaxy Tab S2: 237 × 169 mm display area / 2048
× 1536 pixels), laid flat on the desk. Participants were video-
recorded.
Procedure
All participants performed the 3 manipulation events: Click
and Drag & Drop, Leave on vs. Lift off and Bend vs. Leave
flat. Presentation order was counterbalanced using a Latin
Square. All events involved the flexible version of the 6 T
O-
KENS introduced in [9]: 2 circles, 2 squares, 1 triangle, 1
rectangle.
Event
1
: Click and Drag & Drop. Participants had to perform
2 types of A
CTIONS: Click or Drag. In the Click case, they
had to grab the right token using 3 fingers, put it on a black
cross, and then slide it toward a red circle located 130 mm
away. Once the token was inside the circle, they had to per-
form a “click” on the token by compressing it sideways, and
then release the pressure. Finally, they removed the token
from the surface. In the Drag case, they had to: compress
the token right after having put it on the black cross, keep it
compressed while moving it toward the red circle, and release
the pressure before removing the token from the surface. We
collected data involving sliding movements in 4 main D
IREC-
TIONs: up, down, left, right. The tablet was placed in land-
scape mode for D
IRECTION = {left, right}, and portrait mode
for D
IRECTION = {up, down}, so that the red circle would be
at the same distance from the black cross in all conditions.
Event
2
: Leave on vs. Lift off. Participants also had to move
a token from a black cross to a red circle. However, once in
the circle, participants had to perform one of two A
CTIONS:
< 450ms
Clicking
> 450ms
Dragging
0
1000
2000
0
1000
2000
0.74mm
0
2
-2
-4
d
ref
- d (mm)
Time (ms)
0.74mm
Figure 4. Using Squeeze mode for clicking (left) and dragging (right).
Leave on or Lift off. In the first case, they had to lift their
fingers off the surface but leave the token on it. In the second
case, they had to lift their fingers, taking the token off the sur-
face. We used the same 4 D
IRECTIONs as in Event
1
. We in-
troduced an additional factor, F
INGERCOUNT, to capture the
two different manipulation styles described in [9]: once a to-
ken has been identified with the 3-finger hold, users can keep
manipulating it with 3 fingers, or they can relax their grasp
and manipulate the token with only 2 fingers. Thus, we had
2 F
INGERCOUNT conditions: participants either had to keep
their 3 fingers in contact with the surface all along (3-finger
condition), or they were asked to lift a finger off the surface
after having put the token on the black cross, and to keep it
lifted until the end of the trial (2-finger condition). Failure to
comply in any given trial meant it had to be performed again.
Event
3
: Bend vs. Leave flat. The tablet only displayed a
black cross. Participants had to put the right token on the sur-
face and perform one of two A
CTIONS. In the Bend condition,
they had to bend the token, keeping only their thumb in con-
tact with the surface, and then unbend the token by putting
the other two fingers back on the surface. In the LeaveFlat
condition, they also had to lift two fingers off the tablet, only
keeping the thumb in contact, but without bending the token,
which remained flat on the tablet. They then had to put their
two fingers back on the surface to end the trial.
For each event type, trials are first blocked by A
CTION,
then by D
IRECTION within each ACTION (Event
1
and
Event
2
), and by FINGERCOUNT within each DIRECTION
block (Event
2
). Each condition is replicated 3 times. Block
presentation order is counterbalanced across participants;
trial presentation order within a block is random. The whole
procedure consists of 252 trials (72 + 144 + 36), and lasts
approximately one hour.
RECOGNIZERS
Our main hypothesis was that the micro-movements of inter-
est to us could be observed by looking at the fingers’ traces,
that should move slightly toward, or away from, the token’s
center. To verify this, we analyzed, for all collected touch
traces, the evolution over time of the average distance
d of a
touch point to the centroid of the corresponding multi-touch
sample. In the following, we report the criteria we identified
as the most successful for capturing these micro-movements.
Parameter values (in bold) are determined in the next section.
1. Squeeze: a token is considered squeezed (Figure 4) when:
i ∈{1..|B|},
d
re f
d
i
> d
sqz

where d
re f
is the average distance in millimeters of a touch
point to the centroid of the corresponding multi-touch sample
when users register the token, and B is a buffer containing the
successive values of
d over the last buffer
sqz
milliseconds.
2. On/Off: a token is considered as left on the surface when:
m
end
> m
on_off
where m
end
is the slope
1
of the evolution of d over the
buffer
on_off
milliseconds preceding the instant where the last
finger has been lifted off the surface (count( f ingers)=0).
On the opposite, if m
end
0 at this instant, the token is con-
sidered as lifted off the surface. Figure 5 illustrates the two
cases.
3. Bend: a token is considered as having been bent when:
max(m
be f ore
, m
a f ter
) < 0
where m
be f ore
(resp. m
a f ter
) is the slope of the evolution of d
over the buffer
bend
milliseconds preceding (resp. following)
the instant where only one finger remains in contact with the
surface (count( f ingers)=1) for at least 100ms, as illustrated
in Figure 6. The formula is basically a sign analysis: it checks
whether
d increases or decreases before and after the time
span during which there is one single contact point. We ini-
tially considered analyzing only m
be f ore
to detect when users
enter the bent state, but our tests revealed that this sample
does not carry enough information to discriminate between
bending and leaving flat. This entails that our recognizer con-
siders bent as a discrete event, that gets triggered only once
users have unbent the token.
We couple these criteria with state machines that take the
number of contact points into account, making it very un-
likely that any one event will get confounded with the other
two:
The criterion for squeeze is only evaluated when there are
3 contact points on the surface for at least 200ms. This is
mainly to avoid confusion with cases where users bend the
token, as they tend to compress it when unbending.
The criterion for on/off is only evaluated when the number
of contact points becomes null.
The criterion for bend is only evaluated after a time span of
100ms during which there has been exactly 1 contact point.
RECOGNIZER PARAMETERIZATION
For each of our three micro-movements, we measure the ac-
curacy of our recognizer by running it on data collected for
this micro-movement only. We then test its robustness to false
positives by running it on data collected for the other two.
We use the leave-one-out cross-validation technique to param-
eterize the recognizers: for each participant, we set the param-
eters to values that maximize the overall recognition score for
the 11 other participants. We then report the average score
across all 12 participants (mean, median, standard dev.).
Squeezed mode is recognized in 96.9% (median: 97.9 / std:
3.0) of all trials collected for Event
1
(with d
sqz
[0.74, 0.75]
1
Computed using the Theil-Sen estimator [11].
Time (ms)
Leaving on
40ms
m
end
> 0.0021
40ms
Lifting off
0
50
100
0
50
100
!"#$%&'()&
!%*+%&'()**
!"#$%&'()&
d (mm)
2
m
end
< 0.0021
Figure 5. Leaving a token on the surface (left) or lifting it off (right).
0
2
-2
-4
count(fingers) == 1
160ms
Bending
0
1000
2000
Time (ms)
0
1000
2000
160ms
Leaving flat
m
before
160ms
m
after
d
ref
- d (mm)
m
before
m
after
160ms
> 100ms
count(fingers) == 1
> 100ms
Figure 6. Bending a token (left) or leaving it flat (right).
and buffer
sqz
= 100). It is falsely detected in 1.8% of all trials
for Event
2
, and 2.1% for Event
3
.
States on and off were properly distinguished in 90.1% (me-
dian: 92.4 / std: 5.1)
2
of all trials for Event
2
(with m
on_off
[0.0018, 0.0027] and buffer
on_off
= 40). The distinction
between states on and off also works well for Event
3
, with
only 7.6% of false positives. However, when tested on tri-
als from Event
1
, we observe 43% of false positives. A finer
analysis reveals that the recognizer fails to detect state off
right after leaving mode squeezed, which happens when users
lift the token off while releasing the pressure applied on the
token (
d increases right before count( f ingers) = 0). Mak-
ing tokens flexible thus provides opportunities for performing
micro-movements in general, but has the side-effect of intro-
ducing some ambiguity in this particular case. This is a limita-
tion of our recognizer that we will further investigate. In the
meantime, it can be handled by considering the state where
count( f ingers) = 0 right after having left mode squeezed as
“uncertain”, prompting users for input to resolve the ambigu-
ity.
For E vent
3
, Bent events were detected in 91.1% (median:
91.7 / std: 6.1) of all trials where A
CTION = Bend (with
buffer
bend
[100, 160]). In the remaining 8.9% trials, the
recognizer detected either 0 or at least 2 Bent events (during
the same trial). No Bent event is ever accidentally triggered
for either Event
1
or Event
2
, as the time intervals during which
users have only one finger in contact with the surface are in-
frequent and very short. No Bent event is ever accidentally
triggered, either, when A
CTION = LeaveFlat.
Finally, some indications about the robustness of our flexible-
token design: we used the same set of six tokens throughout
the entire data collection procedure, that consisted of 3024
manipulations by 12 people. No token was broken, or de-
formed.
2
As a side note, we observed a recognition accuracy close to 90%
for on/off states during informal tests using rigid tokens, suggesting
that these micro-movements can also be detected on regular Touch-
Tokens.

Citations
More filters
Proceedings ArticleDOI
11 Oct 2018
TL;DR: The interaction space of MobiLimb is explored, a small 5-DOF serial robotic manipulator attached to a mobile device that overcomes some limitations of mobile devices, and provides rich haptic feedback such as strokes, pat and other tactile stimuli on the hand or the wrist to convey emotions during mediated multimodal communications.
Abstract: In this paper, we explore the interaction space of MobiLimb, a small 5-DOF serial robotic manipulator attached to a mobile device. It (1) overcomes some limitations of mobile devices (static, passive, motionless); (2) preserves their form factor and I/O capabilities; (3) can be easily attached to or removed from the device; (4) offers additional I/O capabilities such as physical deformation and (5) can support various modular elements such as sensors, lights or shells. We illustrate its potential through three classes of applications: As a tool, MobiLimb offers tangible affordances and an expressive controller that can be manipulated to control virtual and physical objects. As a partner, it reacts expressively to users' actions to foster curiosity and engagement or assist users. As a medium, it provides rich haptic feedback such as strokes, pat and other tactile stimuli on the hand or the wrist to convey emotions during mediated multimodal communications.

18 citations


Cites background from "Passive yet Expressive TouchTokens"

  • ...For instance, TouchToken [42] or GaussBricks [40] allow to move, squeeze and stretch tangibles at the surface of the screen, whereas Capstones [9] also provides an extra dimension by detecting stackable elements....

    [...]

Proceedings ArticleDOI
18 Jun 2019
TL;DR: This work-in-progress explores the design space of low-burden in-situ self-reporting by analyzing and mapping context-dependent attention resources, current interactive methods, and associated design requirements.
Abstract: In-situ self-reporting is a widely-used data collection method which offers many benefits in the clinical, psychological and social research fields. However, high capture burden issues have surfaced as in-situ self-reporting expands and diversifies in various studies. Thus, we draw attention to the design space of low-burden in-situ self-reporting. In this work-in-progress, drawing on literature analysis, we explore the design space by analyzing and mapping context-dependent attention resources, current interactive methods, and associated design requirements. In the case study, we further demonstrate the use of the design space to derive low-burden experience sampling solutions. Overall, we stress that reducing in-situ self-report burdens requires research attention, and the design space can help designers make sensible design decisions.

9 citations


Cites methods from "Passive yet Expressive TouchTokens"

  • ...Also, with a Tangible User Interface (TUI), users can change the shape of objects by squeezing or pressing as a way of data input[1, 30]....

    [...]

Proceedings ArticleDOI
14 Feb 2021
TL;DR: In this paper, focused ultrasound is used to transfer power from an ultrasound array commonly used for mid-air haptic feedback and investigate the practical challenges of ultrasonic power transfer (e.g., receiving and rectifying energy from sound waves).
Abstract: Wireless power transfer creates new opportunities for interaction with tangible and wearable devices, by freeing designers from the constraints of an integrated power source. We explore the use of focused ultrasound as a means of transferring power to a distal device, transforming passive props into dynamic active objects. We analyse the ability to transfer power from an ultrasound array commonly used for mid-air haptic feedback and investigate the practical challenges of ultrasonic power transfer (e.g., receiving and rectifying energy from sound waves). We also explore the ability to power electronic components and multimodal actuators such as lights, speakers and motors. Finally, we describe exemplar wearable and tangible device prototypes that are activated by UltraPower, illustrating the potential applications of this novel technology.

8 citations

Proceedings ArticleDOI
29 May 2018
TL;DR: TouchTokenBuilder and TouchTokenTracker are introduced that, taken together, aim at facilitating the development of tailor-made tangible interfaces, showing the strengths and limitations of tangible interfaces with passive tokens.
Abstract: TouchTokens were introduced recently as a means to design low-cost tangible interfaces. The technique consists in recognizing multi-touch patterns associated with specific tokens, and works on any touch-sensitive surface, with passive tokens that can be made out of any material. TouchTokens have so far been limited to a few basic geometrical shapes only, which puts a significant practical limit to how tailored token sets can be. In this article, we introduce TouchTokenBuilder and TouchTokenTracker that, taken together, aim at facilitating the development of tailor-made tangible interfaces. TouchTokenBuilder is an application that assists interface designers in creating token sets using a simple direct-manipulation interface. TouchTokenTracker is a library that enables tracking the tokens' full geometry. We report on experiments with those tools, showing the strengths and limitations of tangible interfaces with passive tokens.

6 citations

Proceedings ArticleDOI
19 Apr 2023
TL;DR: In this article , a design and fabrication pipeline for making mechanical-electronical objects with laser cutting is proposed, which leverage the neodymium magnet's natures of magnetism and conductivity to integrate electronics and mechanical structure joints into prototypes.
Abstract: Laser cutting revolutionizes the creation of personal-fabricated prototypes. These objects can have transformable properties by adopting different materials and be interactive by integrating electronic circuits. However, circuits in laser-cut objects always have limited movements, which refrains laser cutting from achieving interactive prototypes with more complex movable functions like mechanisms. We propose MechCircuit, a design and fabrication pipeline for making mechanical-electronical objects with laser cutting. We leverage the neodymium magnet’s natures of magnetism and conductivity to integrate electronics and mechanical structure joints into prototypes. We conduct the evaluation to explore technological parameters and assess the practical feasibility of the fabrication pipeline. And we organized a user-observing workshop for non-expert users. Through the outcoming prototypes, the result demonstrates the feasibility of MechCircuit as a useful and inspiring prototyping method.
References
More filters
Journal ArticleDOI
TL;DR: In this article, a simple and robust estimator of regression coefficient β based on Kendall's rank correlation tau is studied, where the point estimator is the median of the set of slopes (Yj - Yi )/(tj-ti ) joining pairs of points with ti ≠ ti.
Abstract: The least squares estimator of a regression coefficient β is vulnerable to gross errors and the associated confidence interval is, in addition, sensitive to non-normality of the parent distribution. In this paper, a simple and robust (point as well as interval) estimator of β based on Kendall's [6] rank correlation tau is studied. The point estimator is the median of the set of slopes (Yj - Yi )/(tj-ti ) joining pairs of points with ti ≠ ti , and is unbiased. The confidence interval is also determined by two order statistics of this set of slopes. Various properties of these estimators are studied and compared with those of the least squares and some other nonparametric estimators.

8,409 citations


Additional excerpts

  • ...75] 1Computed using the Theil-Sen estimator [11]....

    [...]

Proceedings ArticleDOI
15 Feb 2007
TL;DR: The reac Table is presented, a musical instrument based on a tabletop interface that exemplifies several of the reasons for which live music performance and HCI in general, and musical instruments and tabletop interfaces in particular, can lead to a fertile two-way cross-pollination that can equally benefit both fields.
Abstract: In recent years we have seen a proliferation of musical tables. Believing that this is not just the result of a tabletop trend, in this paper we first discuss several of the reasons for which live music performance and HCI in general, and musical instruments and tabletop interfaces in particular, can lead to a fertile two-way cross-pollination that can equally benefit both fields. After that, we present the reac Table, a musical instrument based on a tabletop interface that exemplifies several of these potential achievements.

626 citations

Proceedings ArticleDOI
04 Apr 2009
TL;DR: This work proposes to discriminate, among thumb gestures, those it calls MicroRolls, characterized by zero tangential velocity of the skin relative to the screen surface, and shows that at least 16 elemental gestures can be automatically recognized.
Abstract: The input vocabulary for touch-screen interaction on handhelds is dramatically limited, especially when the thumb must be used. To enrich that vocabulary we propose to discriminate, among thumb gestures, those we call MicroRolls, characterized by zero tangential velocity of the skin relative to the screen surface. Combining four categories of thumb gestures, Drags, Swipes, Rubbings and MicroRolls, with other classification dimensions, we show that at least 16 elemental gestures can be automatically recognized. We also report the results of two experiments showing that the roll vs. slide distinction facilitates thumb input in a realistic copy and paste task, relative to existing interaction techniques.

143 citations


"Passive yet Expressive TouchTokens" refers methods in this paper

  • ...We achieve this with a novel design that makes the tokens flexible, and with an analysis of the micro-movements users make when performing these interactions, following an approach similar to the recognizers designed to detect thumb-tip micro-gestures [2, 10]....

    [...]

Proceedings ArticleDOI
05 May 2012
TL;DR: This paper demonstrates three types of tangibles, called CapStones, Zebra Dials and Zebra Sliders, that work with current consumer hardware and investigates what designs may become possible as touchscreen hardware evolves.
Abstract: Recent research proposes augmenting capacitive touch pads with tangible objects, enabling a new generation of mobile applications enhanced with tangible objects, such as game pieces and tangible controllers. In this paper, we extend the concept to capacitive tangibles consisting of multiple parts, such as stackable gaming pieces and tangible widgets with moving parts. We achieve this using a system of wires and connectors inside each block that causes the capacitance of the bottom-most block to reflect the entire assembly. We demonstrate three types of tangibles, called CapStones, Zebra Dials and Zebra Sliders that work with current consumer hardware and investigate what designs may become possible as touchscreen hardware evolves.

132 citations


"Passive yet Expressive TouchTokens" refers background in this paper

  • ...CapStones and ZebraWidgets [3] are capacitive units that can be assembled to configure different conductive circuits, enabling more manipulations with the tangibles that can, for example, be stacked or feature moving parts....

    [...]

Proceedings ArticleDOI
07 May 2011
TL;DR: Two user studies reveal how Facet-Streams unifies visual and tangible expressivity with simplicity in interaction, supports different strategies and collaboration styles, and turns product search into a fun and social experience.
Abstract: We introduce "Facet-Streams", a hybrid interactive surface for co-located collaborative product search on a tabletop. Facet-Streams combines techniques of information visualization with tangible and multi-touch interaction to materialize collaborative search on a tabletop. It harnesses the expressive power of facets and Boolean logic without exposing users to complex formal notations. Two user studies reveal how Facet-Streams unifies visual and tangible expressivity with simplicity in interaction, supports different strategies and collaboration styles, and turns product search into a fun and social experience.

120 citations

Frequently Asked Questions (1)
Q1. What are the contributions in "Passive yet expressive touchtokens" ?

González et al. this paper used a laser-cut lattice hinge to increase the expressive power of touch tokens by making them flexible.