PROJECT TITLE :

Predicting Movie Trailer Viewer's “Like/Dislike” via Learned Shot Editing Patterns

ABSTRACT:

Nowadays, there are a number of movie trailers publicly on the market on social media web site such as YouTube, and many thousands of users have independently indicated whether they like or dislike those trailers. Although it's understandable that there are multiple factors that would influence viewers' like or dislike of the trailer, we aim to address a preference question in this work: Can subjective multimedia options be developed to predict the viewer's preference presented by like (by thumbs-up) or dislike (by thumbs-down) during and when watching movie trailers? We have a tendency to designed and implemented a computational framework that is composed of low-level multimedia feature extraction, feature screening and selection, and classification, and applied it to a assortment of 725 movie trailers. Experimental results demonstrated that, among dozens of multimedia features, the only low-level multimedia feature of shot length variance is extremely predictive of a viewer's “like/dislike” for a large portion of movie trailers. We have a tendency to interpret these findings such that variable shot lengths in a very trailer tend to provide a rhythm that is doubtless to stimulate a viewer's positive preference. This conclusion was additionally proved by the repeatability experiments results using another 600 trailer videos and it was more interpreted by viewers'eye-tracking information.