|"Attractiveness Computing using Artificial Intelligence" |
|Speaker ||: ||Assoc. Prof. Dr. Toshihiko Yamakasi |
|Date ||: ||1 March 2019 (Friday) |
|Time ||: ||13:30 - 15:00 |
|Location ||: ||Faculty of Computer and Informatics Engineering, Ömer Korzay Conference Hall |Seminar Topic:
We have been trying to predict and even enhance the subjective attractiveness of images and videos by using deep neural networks (DNNs). DNNs are now being used not only for image recognition tasks but also for predicting attractiveness of the images such as image aesthetics, face attractiveness scores, etc. Our research focuses on more general and advanced problems by extending such researches: clarifying what contributes to the attractiveness score and how much they contribute to the attractiveness, and enhancing the attractiveness score by DNN-based processing. In our talk, we will introduce some successful cases such as presentation analysis, TV advertisement assessment, hash tag recommendation.
Toshihiko Yamakasi, Associate Professor Department of Information and Communication Engineering, The University of Tokyo