Gazemarks-Gaze-Based Visual Placeholders to Ease Attention Switching Dagmar Kern * Paul Marshall # Albrecht Schmidt * *

Transcription

1 CHI Atlanta -Gaze-Based Visual Placeholders to Ease Attention Switching Dagmar Kern * Paul Marshall # Albrecht Schmidt * * University of Duisburg-Essen # Open University dagmar.kern@uni-due.de, p.marshall@open.ac.uk albrecht.schmidt@acm.org Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

2 Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Validating observations in a formative pilot study Validating observations in a formative pilot study -setup Study task: compare two telephone lists (40 names, one list on paper and one on the screen) 30 participants (volunteers; 10 female, 20 male; aged 23 to 61) How do people compare printed and digital lists? Participants were not aware of the research question No time limit was enforced and time was not measured Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Validating observations in a formative pilot study -results 22 out of 30 used fingers and/or objects to remember their last position on paper list 8 out of 30 used fingers, objects, or cursor on the digital list People use fingers and objects as markers in tasks that require attention switching Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

3 Easing attention switching by providing markers in the digital representation Easing continuation of a task after an interruption Based on implicit interaction requiring no extra effort from the user Idea: detect last gaze position before the attention switch and highlight this area until the gaze returns Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

4 (1) What you look at is what you get: eye movement-based interaction techniques Jacob, R. J What you look at is what you get: eye movement-based interaction techniques. InProceedings of the SIGCHI Conference on Human Factors in Computing Systems: Empowering People(Seattle, Washington, United States, April 01-05, 1990). J. C. Chew and J. Whiteside, Eds. CHI '90. ACM, New York, NY, DOI= Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, (2) Gaze-based interaction for semiautomatic photo cropping (3) Gaze-enhanced scrolling techniques Santella, A., Agrawala, M., DeCarlo, D., Salesin, D., and Cohen, M Gazebased interaction for semi-automatic photo cropping. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems(Montréal, Québec, Canada, April 22-27, 2006). Eds. CHI '06. ACM, New York, NY, DOI= Kumar, M. and Winograd, T Gaze-enhanced scrolling techniques. InProceedings of the 20th Annual ACM Symposium on User interface Software and Technology(Newport, Rhode Island, USA, October 07-10, 2007). UIST '07. ACM, New York, NY, DOI= Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, How to implement? Acquirethe gaze position data (x,y)of the eye tracker in real time Detectwhen the user switches attention between displays Save last gaze position when the user leaves the screen On returning to a display visualize the last gaze position Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

5 Acquire the gaze position data of the eye tracker in real time Hardware Commercial eye tracker TobiiX120 (with a data rate of 120Hz) Software/Middleware EI-Toolkit (component based architecture) Proxy for the Tobii eye-tracker Stream of (x,y) coordinates available to applications over UDP Detect attention switches What do users perceive as the last position they looked at on a screen? How long is the last conscious gaze fixation? How can we detect the perceived last position? How to avoid showing when users blink with their eyes? Conscious gaze = set of fixations/glances at a region within a radius of the first fixation Delay of 600ms before showing the Gazemark Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Visualizing Flag Spotlight Focus area Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

6 Visualizing Focus group with 6 people to choose the visual representation Flag Spotlight Focus area Questionnaire to express a preference Visualizing votedas1 st votedas2 nd votedas3 rd average Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Spotlight Visualizing Easy to find Perceived as precise Not hiding any information Users favored the spotlight visualization Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Setup 2 screen setup (driving simulation setup) 42 Display 8 Display TobiiX120 PC to drive screen and run application Networked over LAN Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

7 Experiment Tasks Map search task on small screen IQ questions on large screen Procedure Find given letter attention switch, solve IQ task find given letter again Design Within-group, each subject completes both task in counterbalanced order Independent variable: visualization (Spotlight vs. no visual aid) Dependent variable: search time (limit 3000 ms) 16 Participants (23 to 52 years old) Dagmar Kern, Paul Marshall, Albrecht Schmidt, Search C Dagmar Kern, Paul Marshall, Albrecht Schmidt, A E F B C G D Which one of the five is least like the other four? A) Dog B) Mouse C) Lion D) Snake E) Elephant Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, What color has the 5 in the circle of around C A B F D E C G Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

8 Results Participants were considerably (about 3 times) faster in searching with with : ms (median) without : ms (median) Comparing search times Strong significant difference (non-parametric Wilcoxon signed-rank test) Qualitative results Preferences for Gazemark Mean value 4.26, standard deviation 0.53, on a scale from 0 (completely senseless) to 5 (very sensible) Reported benefits as perceived by the user: enabling rapid task switch and less attention required Implications on Car UIs Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, Enabling secondary tasks while driving Secondary tasks while driving without without > 2 seconds > 2 seconds eyes on road switching eyes on road switching task task eyes on road switching eyes on road switching task task with with =2 seconds =2 seconds eyes on road task eyes on road task switching switching eyes on road task eyes on road task switching switching Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt, & Future Work can be implemented to support attention switching The time for switching between displays/tasks can be speeded up by a factor of 3 Potential domains Multi-screen/large screen setups In-car information and entertainment systems Switching between multiple documents of which only one is visible at the time (e.g. tabs in a browser) Information terminal in an operating theater Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,

9 Questions? Comments? Visit our websites at: Questions received Penalty for false positives? What is the impact on the timing if the gazemarkis in the wrong place? Do we need to encourage switching while driving? On the implications on driving slide: should there be a switching back time in the diagram (yes but it is the same for both conditions)? What is the impact of the target size (e.g. the size of C)? Dagmar Kern, Paul Marshall, Albrecht Schmidt, Dagmar Kern, Paul Marshall, Albrecht Schmidt,