Shot Scale Classification in Movies by CNN

Abstract

The scale of shot, i.e. the apparent distance of the camera from the main subject of a scene, is one of the main stylistic and narrative functions of filmic products, conveying meaning and inducing the viewer's emotional state while watching movies.
In this work we propose to use convolutional neural networks for the automatic classification of shot scale into Close-, Medium-, or Long-shots.
The development of such a tool will allow to systematically investigate the relationship between the shot scale and the emotional involvement in movies on large movie corpora.
Training and testing are performed on the full filmographies by six different authors (Scorsese, Godard, Tarr, Fellini, Antonioni, and Bergman) for a total number of 120 movies analysed second by second, with results superior to state-of-the-art, thus opening up interesting lines of cross-disciplinary research on the impact of style on the aesthetic and emotional effects on the viewers.

Dataset

The dataset includes the almost complete filmography by six different directors whose styles are consensually considered highly unique and distinguishable in film historiography of author cinema:

Those with * are used for testing.

Results