Related Experiment Videos

Image-type dependent eigen-regions-of-interest define conspicuity operators for predicting human scanpath fixation

Toyomi Fujita1, Claudio M Privitera, Lawrence W Stark

  • 1Neurology and Telerobotics Units, School of Optometry, University of California Berkeley, CA 94720-2020, USA. tfujita@scan.berkeley.edu

Summary

This study reveals how visual perception integrates top-down information and bottom-up conspicuity. A new algorithm predicts human eye movements by analyzing visual conspicuity features from scanpath data.

Related Concept Videos