Gaze-Based Interaction
Darius Miniotas
Dept. of Electronic Systems, VGTU
darius.miniotas@el.vgtu.lt
Gaze-Based Interaction: Why?*
• The only option for disabled
users
• An option for others in hands-
busy situations
• Eye movements
- are extremely fast
- are natural
- require little conscious effort
* Some of the information in these slides is based on the materials presented in a
tutorial at NordiCHI 2004 by Kari-Jouko Räihä, Aulikki Hyrskykari and Päivi Majaranta
Demo from the German Research
Center for AI
• Text 2.0 [2009] – Imagine there were input devices
which could allow text to know if and how it is being read
- how would this change the reading experience?
http://text20.net/node/4
Technological Challenges
• Cost of equipment
- 2000 – 25 000 EUR
- mass production could lower the cost by an order of
magnitude
• Usability of equipment
- remote trackers convenient, but allow only small
movements
- head-mounted trackers more accurate but obtrusive
• Need for calibration
- for every user at the beginning of a tracking session
- often recalibration required during prolonged use
Types of Eye Tracking Applications
• Off-line applications
- visualizing gaze data
- analyzing gaze behavior
- modifying images based on viewing data
• On-line (interactive) applications
- command-based
- attentive
Command-based interaction:
challenges
• Eyes are normally used for observation, not for control
- humans are not used to activating objects just by
looking at them
• Gaze behaves very differently from other ways used for
controlling computers (hands, voice)
- intentional control of eyes is difficult and stressful
- the gaze is easily attracted by external events
- precise control of eyes is difficult
- poorly implemented eye control can be extremely
annoying
“Midas Touch” Problem
• Most of the time the eyes are used for obtaining
information with no intent to initiate commands
• Users are easily afraid of looking at the “eye-active”
objects or areas of the window
• Using eyes for commands requires development of new
forms of interaction
Expanding Targets [CHI 2004]
Selecting Standard-Size Menu Items
[ICMI 2005]
Gaze-Aware Applications
• Command-and-Control
applications
- typing (conventional)
- typing (using gaze gestures)
- drawing
- other
• Multimodal Applications
• Gaze-Contingent Displays
• Attentive Interfaces
Typing by Gaze
• A typical eye typing system has
- an on-screen keyboard
- an eye tracker to record eye
movements
- a computer to analyze gaze
behavior
• To type by gaze the user
- focuses on a letter
- gets feedback from the system
- selects the item in focus
EC Key, a typical keyboard
Compact Keyboard Layouts
Dasher
Demo: http://www.youtube.com/watch?v=0d6yIquOKQ0
Using Gaze Gestures for Typing
Drawing with the Eye [2003]
Other eye-controlled applications
• e-mail
• Internet browsing
• accessing online libraries
• games
• interaction with online virtual communities
EyeScroll [2007]
• gaze-enhanced scrolling allows
for automatic, adaptive scrolling
on content being viewed on the
screen
• supports multiple scrolling modes
depending on the user's
preference and reading style
• users can read the content as is
scrolls smoothly or scrolls once
the user has reached the bottom
of the screen
EyePassword, EyeSaver [2007]
• gaze-based password/pin entry:
prevents shoulder surfing and
does not generate any keyboard
or mouse events – more difficult
to use standard event loggers
• screen saver turned on when a
user looks away from the
screen, off when the user looks
back at the screen
EyePhone [2010]
• developed at Dartmouth
College (USA)
• tracks a person’s eye relative
to a phone’s screen
• users activate applications by
blinking
Demo: http://www.technologyreview.com/computing/25369/
Interactive Applications
• Command-based interaction
- typing (conventional)
- typing (using gaze gestures)
- drawing
• Gaze-aware interfaces
- multimodal input
- gaze-contingent displays
- attentive interfaces
Gaze-Aware Applications
• Command-and-Control
applications
- typing (conventional)
- typing (using gaze gestures)
- drawing
- other
• Multimodal Applications
• Gaze-Contingent Displays
• Attentive Interfaces
Gaze as Mouse Accelerator [1999]
• MAGIC pointing
• Two strategies for warping
- always when the point of gaze
moves (“liberal”)
- only after moving the mouse a
little (“cautious”)
• Empirical results
- liked by the users
- interaction was slightly slowed
down by the cautious strategy, but
the liberal strategy was faster than
using just the mouse
Gaze + Hotkeys [2007]
• performs basic mouse
operation
• reduces / eliminates
dependency on the
mouse for most everyday
tasks such as surfing the
web
• look-press-look-release
action to allow for
increasingly accurate
selection
Demo: http://hci.stanford.edu/research/GUIDe/index.html
Gaze + Speech [2006]
Put-That-There [1982]
• Multimodal input (speech, pointing gestures, gaze)
• Eye gaze used for disambiguation (together with
pointing)
• Demo: http://www.poetv.com/video.php?vid=44316
Gaze-Aware Applications
• Command-and-Control
applications
- typing (conventional)
- typing (using gaze gestures)
- drawing
- other
• Multimodal Applications
• Gaze-Contingent Displays
• Attentive Interfaces
Gaze-Aware Applications
• Command-and-Control
applications
- typing (conventional)
- typing (using gaze gestures)
- drawing
- other
• Multimodal Applications
• Gaze-Contingent Displays
• Attentive Interfaces
iDict [2004]
• automatically detects
irregularities in reading
process
• consults the embedded
dictionaries and provides
assistance
Attentive Videoconferencing [1999]
• Multiparty teleconferencing and
document sharing system
• Images rotate to show gaze
direction (who is talking to whom)
• Document “lightspot” (“look at
this” reference)
PONG: The Attentive Robot [2001]
• A robot that understands and reacts to human presence
and visual communication messages
• Detects when a human walks sufficiently close and then
greets the person verbally and visually by displaying a
smile
• Tries to mimic the user’s facial expressions
Attention Sensors: eyePLIANCES
Eye aRe glasses eyeContact sensor
Light fixture with
eyeContact sensor
Time for a demo: EyeChess

WUD 2010 D.Miniotas - Gaze-Based Interaction

  • 1.
    Gaze-Based Interaction Darius Miniotas Dept.of Electronic Systems, VGTU darius.miniotas@el.vgtu.lt
  • 2.
    Gaze-Based Interaction: Why?* •The only option for disabled users • An option for others in hands- busy situations • Eye movements - are extremely fast - are natural - require little conscious effort * Some of the information in these slides is based on the materials presented in a tutorial at NordiCHI 2004 by Kari-Jouko Räihä, Aulikki Hyrskykari and Päivi Majaranta
  • 3.
    Demo from theGerman Research Center for AI • Text 2.0 [2009] – Imagine there were input devices which could allow text to know if and how it is being read - how would this change the reading experience? http://text20.net/node/4
  • 4.
    Technological Challenges • Costof equipment - 2000 – 25 000 EUR - mass production could lower the cost by an order of magnitude • Usability of equipment - remote trackers convenient, but allow only small movements - head-mounted trackers more accurate but obtrusive • Need for calibration - for every user at the beginning of a tracking session - often recalibration required during prolonged use
  • 5.
    Types of EyeTracking Applications • Off-line applications - visualizing gaze data - analyzing gaze behavior - modifying images based on viewing data • On-line (interactive) applications - command-based - attentive
  • 6.
    Command-based interaction: challenges • Eyesare normally used for observation, not for control - humans are not used to activating objects just by looking at them • Gaze behaves very differently from other ways used for controlling computers (hands, voice) - intentional control of eyes is difficult and stressful - the gaze is easily attracted by external events - precise control of eyes is difficult - poorly implemented eye control can be extremely annoying
  • 7.
    “Midas Touch” Problem •Most of the time the eyes are used for obtaining information with no intent to initiate commands • Users are easily afraid of looking at the “eye-active” objects or areas of the window • Using eyes for commands requires development of new forms of interaction
  • 8.
  • 9.
  • 10.
    Gaze-Aware Applications • Command-and-Control applications -typing (conventional) - typing (using gaze gestures) - drawing - other • Multimodal Applications • Gaze-Contingent Displays • Attentive Interfaces
  • 11.
    Typing by Gaze •A typical eye typing system has - an on-screen keyboard - an eye tracker to record eye movements - a computer to analyze gaze behavior • To type by gaze the user - focuses on a letter - gets feedback from the system - selects the item in focus EC Key, a typical keyboard
  • 12.
  • 13.
  • 14.
  • 15.
    Drawing with theEye [2003]
  • 16.
    Other eye-controlled applications •e-mail • Internet browsing • accessing online libraries • games • interaction with online virtual communities
  • 17.
    EyeScroll [2007] • gaze-enhancedscrolling allows for automatic, adaptive scrolling on content being viewed on the screen • supports multiple scrolling modes depending on the user's preference and reading style • users can read the content as is scrolls smoothly or scrolls once the user has reached the bottom of the screen
  • 18.
    EyePassword, EyeSaver [2007] •gaze-based password/pin entry: prevents shoulder surfing and does not generate any keyboard or mouse events – more difficult to use standard event loggers • screen saver turned on when a user looks away from the screen, off when the user looks back at the screen
  • 19.
    EyePhone [2010] • developedat Dartmouth College (USA) • tracks a person’s eye relative to a phone’s screen • users activate applications by blinking Demo: http://www.technologyreview.com/computing/25369/
  • 20.
    Interactive Applications • Command-basedinteraction - typing (conventional) - typing (using gaze gestures) - drawing • Gaze-aware interfaces - multimodal input - gaze-contingent displays - attentive interfaces
  • 21.
    Gaze-Aware Applications • Command-and-Control applications -typing (conventional) - typing (using gaze gestures) - drawing - other • Multimodal Applications • Gaze-Contingent Displays • Attentive Interfaces
  • 22.
    Gaze as MouseAccelerator [1999] • MAGIC pointing • Two strategies for warping - always when the point of gaze moves (“liberal”) - only after moving the mouse a little (“cautious”) • Empirical results - liked by the users - interaction was slightly slowed down by the cautious strategy, but the liberal strategy was faster than using just the mouse
  • 23.
    Gaze + Hotkeys[2007] • performs basic mouse operation • reduces / eliminates dependency on the mouse for most everyday tasks such as surfing the web • look-press-look-release action to allow for increasingly accurate selection Demo: http://hci.stanford.edu/research/GUIDe/index.html
  • 24.
  • 25.
    Put-That-There [1982] • Multimodalinput (speech, pointing gestures, gaze) • Eye gaze used for disambiguation (together with pointing) • Demo: http://www.poetv.com/video.php?vid=44316
  • 26.
    Gaze-Aware Applications • Command-and-Control applications -typing (conventional) - typing (using gaze gestures) - drawing - other • Multimodal Applications • Gaze-Contingent Displays • Attentive Interfaces
  • 27.
    Gaze-Aware Applications • Command-and-Control applications -typing (conventional) - typing (using gaze gestures) - drawing - other • Multimodal Applications • Gaze-Contingent Displays • Attentive Interfaces
  • 28.
    iDict [2004] • automaticallydetects irregularities in reading process • consults the embedded dictionaries and provides assistance
  • 29.
    Attentive Videoconferencing [1999] •Multiparty teleconferencing and document sharing system • Images rotate to show gaze direction (who is talking to whom) • Document “lightspot” (“look at this” reference)
  • 30.
    PONG: The AttentiveRobot [2001] • A robot that understands and reacts to human presence and visual communication messages • Detects when a human walks sufficiently close and then greets the person verbally and visually by displaying a smile • Tries to mimic the user’s facial expressions
  • 31.
    Attention Sensors: eyePLIANCES EyeaRe glasses eyeContact sensor Light fixture with eyeContact sensor
  • 32.
    Time for ademo: EyeChess