MovieCLIP: Visual Scene Recognition in Movies
This addresses the problem of limited scene recognition datasets for movies, enabling better analysis of longform media, though it is incremental by building on existing weak supervision methods.
The authors tackled visual scene recognition in movies by automatically curating a taxonomy of 179 scene labels and using CLIP to weakly label 1.12 million shots, showing that models pretrained on this dataset improve downstream tasks like scene and genre classification.
Longform media such as movies have complex narrative structures, with events spanning a rich variety of ambient visual scenes. Domain specific challenges associated with visual scenes in movies include transitions, person coverage, and a wide array of real-life and fictional scenarios. Existing visual scene datasets in movies have limited taxonomies and don't consider the visual scene transition within movie clips. In this work, we address the problem of visual scene recognition in movies by first automatically curating a new and extensive movie-centric taxonomy of 179 scene labels derived from movie scripts and auxiliary web-based video datasets. Instead of manual annotations which can be expensive, we use CLIP to weakly label 1.12 million shots from 32K movie clips based on our proposed taxonomy. We provide baseline visual models trained on the weakly labeled dataset called MovieCLIP and evaluate them on an independent dataset verified by human raters. We show that leveraging features from models pretrained on MovieCLIP benefits downstream tasks such as multi-label scene and genre classification of web videos and movie trailers.