Accessibility of Brainstorming Sessions for Blind - IWF

Accessibility of Brainstorming Sessions for Blind People
Andreas Kunz1, Klaus Miesenberger2, Max Mühlhäuser3, Ali Alavi1,
Stephan Pölzer2, Daniel Pöll2, Peter Heumader2, and Dirk Schnelle-Walka3
Swiss Federal Institute of Technology, Zurich, Switzerland
Institut Integriert Studieren, Johannes Kepler University, Linz, Austria
Technische Universität Darmstadt, Germany
Abstract. Today, research focuses on the accessibility of explicit information
for blind users. This gives only partly access to the information flow in brainstorming sessions, since non-verbal communication is not supported. Advances
in ICT however allow capturing implicit information like hand gestures as
important part of non-verbal communication. Thus, we describe a system that
al-lows integrating blind people into a brainstorming session using a mind map.
Keywords: Accessibility, Mind map, Non-verbal Communication Elements.
In brainstormings, coordinative, collaborative and communicative elements exist [1]
in the ‘level of artifacts’ and the ‘level of non-verbal communication’ (Fig. 1).
Fig. 1. Brainstorming integrates non-verbal communication and artifact level elements [2]
Within these levels, a number of challenges arise when blind and sighted persons
collaborate. The challenges come from fundamentally different (and often incompatible) ways of perception and expression that cannot be easily overcome. This could
bring blind people in an unwanted role, which could result in an unintended or unconscious exclusion from such a brainstorming session. In case of a brainstorming session
using a mind map, the origin of the problems is twofold: they could either come from
K. Miesenberger et al. (Eds.): ICCHP 2014, Part I, LNCS 8547, pp. 237–244, 2014.
© Springer International Publishing Switzerland 2014
A. Kunz et al.
the generated elements on the artifact level, or from the non-verbal communication
level. Both levels are highly interdependent, since many times non-verbal communication elements, such as deictic gestures, refer to elements on the artifact level.
Level of Artifacts
Although blind and sighted people can access the same digital information of documents by rendering it on screen/print or in Braille/audio, the way of perception leads
to considerable differences in accessing or generating this information. The visual
channel allows for fast and parallel perception of information, while Braille or audio
could be compared to a serial perception. This difference is in particular salient for the
perception of graphical elements like those of a mind map. Hence, blind people need
more time for sensing the content of artifacts. This becomes even more critical when
the artifacts are edited during the meeting. Moreover, jointly working on artifacts
requires that each participant is aware about the state and the content of the artifacts.
Due to these limitations, current tools and technologies cannot support a joint group
of sighted and blind people to create and edit artifacts in a brainstorming meeting.
Level of Non-verbal Communication
A large amount of information during a meeting is transferred non-verbally by gestures, postures, facial expressions, etc., which rely on the visual channel only. Consequently, blind people are excluded from this information. Not only gestures and facial
expressions are required for establishing interpersonal relationship [3], but are also
important to coordinate discussions during a brainstorming, e.g. for turn-taking [4,5].
As a workaround for engaging blind people in such meetings, the implicit non-verbal
communication is made explicit by verbalizing it. This evokes new problems, since
non-verbal communication is parallel and thus has to be serialized in order to communicate it verbally. Moreover, non-verbal communication is done unconsciously,
but it has to be consciously translated or articulated by the sighted participants. Both
problems mentioned in the above slow down the whole brainstorming session.
The mentioned problems show the difficulties for the participants of a meeting to
transfer information and to synchronize the individual mental models. Meetings are
significantly more complicated as they have to be consciously prepared and conducted. This imposes a higher cognitive load on the participants. Consequently, such
meetings tend to progress more slowly, are less effective and less intuitive. Finally,
some activities are even impossible, such as intensely working on a shared document.
Related Work
For the chosen scenario of integrating blind people into a collocated mind map brainstorming session, we need to study the research literature on two different levels: the
artifact level, and the non-verbal communication level.
Artifact Level
A vast body of research in the field of HCI (Human Computer Interaction) tackles the
problem of more intuitive interaction with digital content. A major approach relies on
Accessibility of Brainstorming Sessions for Blind People
natural user interfaces (NUIs), which enhance (or in some case, replace) a graphical
user interface (GUI), buy providing technologies such as touchscreens, motions trackers, and tangible user interfaces (TUIs) [6,7,8,9] . While these interfaces are helpful
for sighted persons to access and alter artifacts in an intuitive way, they support the
blind users in a very limited manner.
Window Eyes [10] could output textual information to Braille displays, or audio
speakers. A lot of research has been done for making graphical information accessible
to the blind, both in terms of providing equivalent descriptive text alternatives [11], or
using tactile and/or audio methods [12,13,14,15]. All these approaches focus on linear
text and static graphics, but not on dynamically changing artifacts such as a mind
map. This requires the development of systems for tracking, analyzing and displaying
information and thereby tackling the problem of information overflow. Only a few
approaches towards access to dynamic content exist [16,17]. Here, the user interacts
with a tangible and sonic representation of a line chart. In another approach, Ferres et
al. [18] translate the chart into a textual representation, which can then be accessed by
conventional reading devices such as Braille displays.
However, these approaches do not investigate the dynamics of changing graphics
in real-time, and they also do not include non-verbal communication elements.
Non-verbal Communication Level
While meeting support for blind people is still a largely underrepresented research
area as stated by Winberg et al. [19], distributed meetings for sighted persons are well
addressed in literature. Providing awareness of eye-contacts or of other gestures such
as pointing or nodding supports the social presence and coordinates the teamwork [5],
[4], [20]. Work presented by Ba et al. [21] provides a proximity approach to identify
the visual focus of attention of a user. Morency et al. [22] show that nodding and
other head movements can be successfully captured by visual tracking systems.
Deictic gestures bridge between the level of non-verbal communication and the
level of artifacts. They are used by sighted people to easily refer to artifacts during a
meeting. Within the very little work conducted for blind people, Oliveira et al. [23]
address how to translate deictic gestures made on a whiteboard containing static content into a haptic representation. Thereby, the blind person wears a glove which
guides his hand to the position on the artifact level the sighted user is pointing to.
However, there is no work for deictic gestures on dynamic content.
To summarize, the integration of blind people into brainstorming sessions is still
very limited. This is mainly because of the following reasons:
• Deictic gestures are not reliably captured and assigned to artifacts on the interactive surface.
• Translation of the "parallel" content of the artifact level to the serial perception of
the blind user and vice versa is not well investigated.
• Lack of interfaces for blind users to interact with the artifacts
This paper thus introduces a system that allows integrating blind users in a mind map
brainstorming meeting, which addresses the challenges mentioned in the above.
A. Kunz et al.
To address the above challenges, we realized the concept of an automatic real-time
translation as core of the overall system. This translation process is threefold: (1)
capturing artifacts and non-verbal communication elements with dedicated sensors;
(2) semantic modeling of artifacts and activities; and (3) making the information
available to all participants in an appropriate and accessible representation (Fig. 2).
Fig. 2. General overview of the realized system
The system is realized in a multi-layer architecture, consisting of 4 layers.
Persistence Layer
The persistence layer only contains the mind map model. Although many graph-based
mind map tools exist, we implemented our mind map in a tree structure. Within a tree
structure, every element n the mind map represents a node in the tree model, thus
having a unique path to the root. This path is made accessible to the blind users. In
comparison to alternative screen exploring techniques for blind users as presented by
Kane et al. [24], a tree structure can be easily navigated, interpreted and manipulated
by a blind user in a way she is already familiar with.
Controller Layer
The Model Update Controller updates the persistence layer. The controller will also
report changes of the model to the application layer. The model update controller also
provides information to the semantic reasoner in order to properly infer the relation
between deictic and other pointing gestures and artifacts.
Application Layer
The main components of the application layer are the mind map software, the reasoner, and an accessible mind map (blind user interface).
of Brainstorming Sessions for Blind People
The reasoner receives th
he tracking information from the sensors and compares the
calculated pointing directio
on intersection point with the positions of the artifactss on
the screen in order to high
hlight the corresponding component. This informationn is
given to the mind map and
d the accessible mind map applications, where the corrresponding object is highlighteed.
The mind map applicatio
on is shown in Fig. 3. It currently allows adding and deeleting elements, adding text to
o an element and freely moving it around. It is also possiible
to rotate an element, which is important since the sighted persons sit around the tabble.
Fig. 3.. Collaborative mind map Editor “CoME“
Sighted persons can mo
ove, zoom, and rotate objects by using the touch innput
capabilities of Microsoft PiixelSense tabletop computer. All structure and contentt information is transferred to the blind user interface (see Fig. 4). This is in principlle a
serialized representation of the parallel information in the mind map. When the m
map is updated, the structurre of the tree in the blind user interface is altered. The innterface also offers possibilitiess to modify the mind map on the table, such as adding, m
modifying, or deleting a node in
n the tree. Moreover, cut and paste functionalities are avvailable, as well as a search fun
nction or a button for further details of a node. Any moddification will immediately be transferred
to the mind map visualization on the PixelSennse.
Fig. 4. Blind user interface
and blind user testing the blind user interface
The blind user interfacee also contains a tree structure representing the contentt of
the mind map. The blind usser can browse this tree in a way he is already familiar w
from other environments. The
T interface also includes a list view with the mind maap’s
history of modifications. This
allows the blind user to view modifications also inn an
asynchronous way. Finally,, a search functionality allows searching the history as w
as the tree. A message box will further notify the blind user when changes are made to
the mind map or when pointing gestures highlight a certain artifact (node). All the
elements can be accessed by
y the blind user with the Braille display (see Fig. 4).
A. Kunz et al.
Input/Output Layer
As described above, artifacts as well as NVCs have to be captured by the system. For
our realized prototype, we choose a scenario (but are not limited to) with three sighted
users and one blind user who gather around PixelSense as shown in Fig. 5.
Fig. 5. Overall physical setup
Since PixelSense is an interactive screen, any touch input for modifying the artifacts can be sensed, while text is currently still entered via the keyboard. For detecting
deictic gestures as the most important representative for NVCs, three LEAP Motion
sensors are used. They are placed on the PixelSense’s frame to detect deictic gestures
from the sighted users (see Fig. 5). The sensors are oriented in such a way that the
lower boundary of the sensor’s field of view is parallel to the table’s surface.
Preliminary User Tests
The blind user interface together with the mind map editor CoME were evaluated in
first trial runs. The blind users accessed the interface with a Braille display which was
connected to a screen reader. Using different screen readers, the blind user interface
was in general accessible, although different readers showed different performance.
In general, the blind users appreciated to present mind maps to different user
groups with different adapted views. It was further seen as a good solution for the
synchronization between blind and sighted users within a mind map session that there
is an alert system which informs the blind user about any modification done on the
mind map. Also the history of the blind user interface was very much appreciated,
since it could be used if blind users missed some changes made in the mind map.
The expected irritations of sighted users by suddenly appearing nodes in the mind
map – being entered via the blind user interface – were not confirmed during our first
trail runs. The reason for this could be found in the fact that there was still an audio
synchronization between all the team members, from which cues about the next possible interaction of the blind user could be derived.
Accessibility of Brainstorming Sessions for Blind People
Conclusion and Outlook
We introduced a system that supports a mind map brainstorming meeting between
blind and sighted users. The system gives access for blind users to the generated
arti-facts, and also captures and transfers non-verbal communication elements. This
allows a deep integration of a blind user into a brainstorming meeting.
Our work will continue with in-depth user studies with the built system. We will
then continue to integrate other NVC elements into the system, such as nodding or
shrugging. For integrating these NVC elements, we will extend our current system by
additional sensors. Moreover, more sophisticated filtering algorithms will be developed to avoid false interpretations and thus wrong notifications to the blind user.
Acknowledgments. This work was done in collaboration of ETH Zurich, TU
Darmstadt, and JKU Linz and was funded under the DACH umbrella under the number CR21I2L_138601.
1. Ellis, C.A., Gibbs, S.J., Rein, G.: Groupware: Some Issues and Experiences. Communications of the ACM 34, 39–58 (1991)
2. Kunz, A., Alavi, A., Sinn, P.: Integrating Pointing Gesture Detection for Enhancing Brainstorming Meetings Using Kinect and PixelSense. In: Proc. of the 8th International Conference on Digital Enterprise Technology, ETH-Zürich, Zürich, Switzerland, pp. 1–8 (2014)
3. von Thun, F.S.: Miteinander reden 1 – Störungen und Klärungen. Allgemeine Psychologie
der Kommunikation. rororo, Reinbek, Germany (1998)
4. Kendon, A.: Some functions of gaze-direction on social interaction. Acta Psychologica 26,
22–63 (1967)
5. Argyle, M., Cook, M.: Gaze & Mutual Gaze. Cambridge University Press, New York
6. Ishii, H., Ullmer, B.: Tangible Bits: Toward Seamless Interfaces between People, Bits and
Atoms. In: Proc. of the SIGCHI Conference on Human Factors in Computing Systems
(CHI 1997), pp. 234–241. ACM, New York (1997)
7. Fitzmaurice, G., Ishii, H., Buxton, W.: Bricks: Laying the Foundation for Graspable User
Interfaces. In: Proceedings of the SIGCHI Conference on Human Factors in Computing
Systems (CHI 1995), pp. 442–449. ACM, New York (1995)
8. Ganser, C., Steinemann, A., Kunz, A.: InfrActables: Supporting Collocated Groupwork by
Combining Pen-based and Tangible Interaction. In: Proceedings of IEEE Tabletop 2007,
pp. 1–2. IEEE, Washington (2007)
9. Hofer, R., Kunz, A.: TNT: Touch ‘n’ Tangibles on LC-displays. In: Natkin, S., Dupire, J.
(eds.) ICEC 2009. LNCS, vol. 5709, pp. 222–227. Springer, Heidelberg (2009)
10. Window Eyes by GW Micro,
11. W3C/WAI Guidelines,
12. Yayl, L.: Huseby Zoom Maps: A Design Methodology for Tactile Graphics. Journal of
Visual Impairment & Blindness 3(5), 270–276 (2009)
A. Kunz et al.
13. Manshad, M.S., Manshad, A.S.: Multimodal Vision Glove for Touchscreens. In: Proceedings of the 10th International ACM SIGACCESS Conference on Computers and Accessibility, pp. 251–252. ACM, New York (2008)
14. King, A., Blenkhorn, P., Crombie, D., Dijkstra, S., Evans, G., Wood, J.: Presenting
UML Software Engineering Diagrams to Blind People. In: Miesenberger, K., Klaus, J.,
Zagler, W.L., Burger, D. (eds.) ICCHP 2004. LNCS, vol. 3118, pp. 522–529. Springer,
Heidelberg (2004)
15. Petrie, H., et al.: TeDUB: A System for Presenting and Exploring Technical Drawings for
Blind People. In: Miesenberger, K., Klaus, J., Zagler, W.L. (eds.) ICCHP 2002. LNCS,
vol. 2398, pp. 537–539. Springer, Heidelberg (2002)
16. Petit, G., Dufresne, A., Levesque, V., Hayward, V., Trudeau, N.: Refreshable Tactile
Graphics Applied to Schoolbook Illustrations for Students with Visual Impairments. In:
Proceedings of the 10th International ACM SIGACCESS Conference on Computers and
Accessibility (Assets 2008), pp. 89–96. ACM, New York (2008)
17. Landua, S., Welss, L.: Merging Table Sensory Input and Audio Data by Means of
the Talking Tactile Tablet. In: Proceedings of Eurohaptics, pp. 414–418. IEEE Computer
Society, Washington (2003)
18. Ferres, L., Verkhogliad, P., Lindgaard, G., Boucher, L., Chretien, A., Lachance, M.:
Improving Accessibility to Statistical Graphs: The iGraph-Lite System. In: Proceedings of
the 9th International SIGACCESS Conference on Computers and Accessibility (Assets
2007), pp. 67–74. ACM, New York (2007)
19. Winberg, F., Bowers, J.: Assembling the Senses: Towards the Design of Cooperative Interfaces for Visually Impaired Users. In: Proceedings of the 2004 ACM Conference on Computer Supported Cooperative Work (CSCW 2004), pp. 332–341. ACM, New York (2004)
20. Kunz, A., Nescher, T., Küchler, M.: CollaBoard: A Novel Interactive Electronic Whiteboard for Remote Collaboration with People on Content. In: Proceedings of the International Conference on Cyberworlds (CW 2010), pp. 430–437. IEEE, Washington (2010)
21. Ba, S.O., Odobez, J.-M.: A Study on Visual Focus of Attention Recognition from Head
Pose in a Meeting Room. In: Renals, S., Bengio, S., Fiscus, J.G. (eds.) MLMI 2006.
LNCS, vol. 4299, pp. 75–87. Springer, Heidelberg (2006)
22. Morency, L.-P., Quattoni, A., Darrell, T.: Latent-Dynamic Discriminative Models for Continuous Gesture Recognition. In: Proceedings of the 2007 IEEE Conference on Computer
Vision and Pattern Recognition (CVPR 2007), pp. 1–8. IEEE, Washington (2007)
23. Oliveira, F., Cowan, H., Faug, B., Quek, F.: Enabling Multimodal Discourse for the Blind.
In: Proceedings of the International Conference on Multimodal Interfaces and the Workshop on Machine Learning for Multimodal Interaction (ICMI-MLMI 2010), vol. (18).
ACM, New York (2010)
24. Kane, S.K., Morris, M.R., Perkins, A.Z., Wigdor, D., Ladner, R.E., Wobbrock, J.O.:
Access Overlays: Improving Non-visual Access to Large Touch screens for Blind Users.
In: Proceedings of the 24th Annual ACM Symposium on User Interface Software and
Technology (UIST 2011), pp. 273–282. ACM, New York (2011)