Automatic System to Remove Unpleasant Images Detected by Pupil-Size Changes
An automatic filtering system to classify individual unpleasant emotions represented by pupil-size changes was proposed. The support vector machines classifier was applied to single-trial data of pupil size and indicated the possibility of the correct judgment of individually unpleasant states immediately after looking at emotional pictures. The framework was then constructed to automatically filter the similar unpleasant information from a picture database, applying the bag of features scheme to search for similar images.
Keywords: Pupil size, Emotion, Database search, Support vector machines
Download Full-Text
ABOUT THE AUTHOR
Koji Kashihara
The University of Tokushima
Koji Kashihara
The University of Tokushima