In this paper, we aim to estimate the Winner of world-wide film festival fromthe exhibited movie poster. The task is an extremely challenging because theestimation must be done with only an exhibited movie poster, without any filmratings and box-office takings. In order to tackle this problem, we havecreated a new database which is consist of all movie posters included in thefour biggest film festivals. The movie poster database (MPDB) contains historicmovies over 80 years which are nominated a movie award at each year. We apply acouple of feature types, namely hand-craft, mid-level and deep feature toextract various information from a movie poster. Our experiments showedsuggestive knowledge, for example, the Academy award estimation can be betterrate with a color feature and a facial emotion feature generally performs goodrate on the MPDB. The paper may suggest a possibility of modeling human tastefor a movie recommendation.
展开▼
