Sie sind hier

Perception-based Attention Analysis

Eye tracking data can reveal information about the attention of a user on the computer screen. This is of high interest for reading studies, advertisements, personalization or usability assessment. In order to perform attention analysis, one has to extract the observed content. As of now, this is performed via video recording of the complete session, screenshot generation or extraction of the structural content, e.g., DOM tree of a Web page.

In this thesis, you would work on a novel approach of fixation-centered screenshots, which we call focus path. The focus path might allow for a simpler procedure to process data than a long video recording, a more complete representation than a single screenshot and it would be independent from any structural information, which means it would work for various kind of stimuli. The idea is as following:
1. Store pixel data from the region of focal vision along the estimated gaze during the user operates the system
2. Stitch stored pixel data according to spatial and visual information (e.g., match consecutive pixel data to compensate for scrolling)
3. Attempt (i) detection of text within stitched pixel data, (ii) detection of changes in the presented stimulus (e.g., menu opens) and (iii) clustering of the focus of multiple users.

You should review the following literature before applying for the thesis and reach out for more related work during the thesis process:
1. Overview about gaze data visualization approaches [1]
2: System to analyze attention on Web pages using extraction of Web page content [2]
3. Studying reading behavior on Web pages [3, Chapter 6]
4. Visualization of gaze data with so-called gaze strips. This is very related to the suggested approach. It stores pixel data from the region of focal vision and presents it as thumbnails in a timeline [4]

Prerequisites
Experience in programming. Preferably in C++, Python and / or Java. Please provide programming references, e.g., personal projects, when you apply for this thesis.

References
[1] Blascheck, T., Kurzhals, K., Raschke, M., Burch, M., Weiskopf, D. and Ertl, T. (2017): Visualization of Eye Tracking Data: A Taxonomy and Survey. Computer Graphics Forum, 36: 260–284. doi:10.1111/cgf.13079
[2] David Beymer and Daniel M. Russell (2005): WebGazeAnalyzer: a system for capturing and analyzing web reading behavior using eye gaze. In CHI '05 Extended Abstracts on Human Factors in Computing Systems (CHI EA '05). ACM, New York, NY, USA, 1913-1916. DOI=http://dx.doi.org/10.1145/1056808.1057055
[3] Drewes, Heiko (2010): Eye Gaze Tracking for Human Computer Interaction. Dissertation, LMU München: Faculty of Mathematics, Computer Science and Statistics
[4] K. Kurzhals, M. Hlawatsch, F. Heimerl, M. Burch, T. Ertl and D. Weiskopf, "Gaze Stripes: Image-Based Visualization of Eye Tracking Data," in IEEE Transactions on Visualization and Computer Graphics, vol. 22, no. 1, pp. 1005-1014, Jan. 31 2016. doi: 10.1109/TVCG.2015.2468091

Betreuer: 
Studienart: 
Master
Ausschreibungsdatum: 
2018