Time series analysis
may reveal important insight about the mechanism driving eye-movements.

Eye Tracking Experiments

Deborah J. Aks

 
tststststs

Attention typically coincides w/ central vision where the image is much clearer compared to peripheral vision. 
Image blurring is one tool we can use to help disambiguate confound of attention & low level image resolution.
  search for:

 

 

Gaze contingent blurring
helps distinguish attention from low-level vision.i.e., Is search performance due to attentional focus or greater image clarity/contrast?


central -->
-vs-
<--peripheral

blurring.

These types of manipulations permit us to distingish the role of attention in search performance.





 

Size & clarity manipulations further distinguish low-level from attentional influences

Search efficiency

evaluated with manipulations of the # of items in the display

search for:

 

letter
'M'
   
Eye-movement time series
   
   

Demanding search tasks produce complex search patterns  in the time series of eye-movements


Are there any patterns in these time series or is this variability just random noise?



Any fractal
patterns?

  
   
How much vaiability is due to instrument noise?

Distinguishing human vs instrument noise



tracking robotic eye mvmts

 
Satellite imagery
 


Many real world tasks involve visual search & produce complex search patterns

 

Locating land- marks often produces complicated patterns


Using these same eye-tracking tasks we can also learn about effective features for representing landmarks
  What is the impact of resolution in satellite images?
-------->
Earlier version of Experiment overview file

<-- search for spider web in noisy background forest scene.

 

http://aks.rutgers.edu/
updated 6/16/06