MovieCLIP: Visual Scene Recognition in Movies

TL;DR

This paper introduces MovieCLIP, a large-scale dataset and model for recognizing complex visual scenes in movies, leveraging weak supervision from CLIP to improve scene understanding and downstream tasks.

Contribution

The work creates a new extensive movie-centric scene taxonomy and a weakly labeled dataset using CLIP, enabling improved visual scene recognition in movies.

Findings

Weakly labeled dataset of 1.12 million shots from 32K movie clips.

Baseline models trained on MovieCLIP outperform previous methods.

Features from MovieCLIP enhance downstream scene and genre classification.

Abstract

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.