http://infolab.stanford.edu/~wangz/project/imsearch/Aesthetics/ICIP08/datta.pdf
Photo.net
DPChallenge
Terragalleria
Alipr
The authors of the study collected data from these four websites that let users rate the quality and aesthetics of photographs. They applied this data to aesthetic algorithms. Three questions were used to accumulate the data:
Score Prediction: Rating the photo 1 to 10, or 1 to 7
Class Prediction: High or Low, Professional or Snapshot
Emotion Prediction: Amusing, scary, boring, pleasing, irritating
No comments:
Post a Comment