Measuring Visual Attention

This project is involved with estimating visual attention and its derivatives to better understand consumer behavior.

• How can we measure the distribution of visual attention in the shopping scenario?
  
• If a customer conducts a targeted product search is attention distributed differently?
  
• Does visual clutter influence the speed and efficiency search tasks?
  
• Can we use eye tracking data to improve our model visual attention?
  
• What are the features we can use. Visual: shape, color, orientation, size. Non-visual: ?
  
• How can we combine low-level saliency with top-down target appearance features?
  
• Can we effectively predict the distribution of visual attention in an unknown scene?
  

For a static scene the only type of visual processes taking place in observer can be described as a sequence of fixations and saccades. These can be computed from an array of gazepoints collected by the eye-tracking hardware. (Salvucci and Goldberg) To detect fixation points we used a robust and accurate algorithm based on dispersion threshold identification - I-DT. (see Widdel, Wedel and Pieters). Our I-DT fixation detection method described above yeilds mean fixation durations of ~307 ms which is consistent with the previous research (Hooge and Erkelens). Typical estimates of attentional shift time in visual search experiments are on the order of 50 ms (Wolfe, 1998).

In our visual attention study we use eye-tracking results as an objective measure of visual attention. These are further used as a training set in a supervised learning process that improved the precision of our predictive models.

Our predictive model uses low level saliency indicators (color, texture) with a combination of hi-level saliency driving factors for targeted search tasks. These features are contracted based on the shape and appearnce of the target search object or a group of objects. Combining top-down and bottom-up saliency can predict the location of up to 70-80 percent of the fixations correctly.

• van der Lans, R.; Pieters, R. & Wedel, M. (2008), 'Competitive Brand Salience', Marketing Science.
• Wolfe, J. M.; Horowitz, T. S. & Anonymous (2004), 'What attributes guide the deployment of visual attention and how do they do it?', Nat Rev Neurosci 5, 495--501.
• Torralba, A.; Oliva, A.; Castelhano, M. S. & Henderson, J. M. (2006), 'Contextual guidance of eye movements and attention in real-world scenes: the role of global features in object search.', Psychol Rev 113(4), 766--786.
• Elazary, L. & Itti, L. (2008), 'Interesting objects are visually salient', Journal of Vision 8(3:3), 1-15.

My Visual Attention Bibliography (in separate window)