Long Story Short: Story-level Video Understanding from 20K Short Films

TL;DR

This paper introduces SF20K, a large dataset of 20,143 amateur films designed to advance long-term video understanding and reasoning, and demonstrates that instruction tuning on this dataset enhances model performance.

Contribution

The paper presents SF20K, the largest publicly available long-form movie dataset, enabling new long-term video tasks and addressing data leakage issues in existing datasets.

Findings

SF20K enables long-term video reasoning tasks.

Recent VLMs perform well on SF20K.

Instruction tuning improves model performance significantly.

Abstract

Recent developments in vision-language models have significantly advanced video understanding. Existing datasets and tasks, however, have notable limitations. Most datasets are confined to short videos with limited events and narrow narratives. For example, datasets with instructional and egocentric videos often depict activities of one person in a single scene. Although existing movie datasets offer richer content, they are often limited to short-term tasks, lack publicly available videos, and frequently encounter data leakage issues given the use of subtitles and other information about commercial movies during LLM pretraining. To address the above limitations, we propose Short-Films 20K (SF20K), the largest publicly available movie dataset. SF20K is composed of 20,143 amateur films and offers long-term video tasks in the form of multiple-choice and open-ended question answering. Our extensive analysis of SF20K reveals minimal data leakage, emphasizes the need for long-term reasoning, and demonstrates the strong performance of recent VLMs. Finally, we show that instruction tuning on the SF20K-Train set substantially improves model performance, paving the way for future progress in long-term video understanding.