TGIF: A New Dataset and Benchmark on Animated GIF Description

TL;DR

This paper introduces TGIF, a large dataset of animated GIFs with natural language descriptions, serving as a benchmark for GIF understanding and captioning, and demonstrates baseline models and their potential for related tasks.

Contribution

The paper presents TGIF, a new high-quality dataset for animated GIF captioning, along with validation methods, statistical analysis, baseline models, and insights into transfer learning for movie description.

Findings

High correlation between visual content and descriptions in TGIF.

Baseline models achieve promising results on GIF captioning.

Fine-tuned models improve automatic movie description.

Abstract

With the recent popularity of animated GIFs on social media, there is need for ways to index them with rich metadata. To advance research on animated GIF understanding, we collected a new dataset, Tumblr GIF (TGIF), with 100K animated GIFs from Tumblr and 120K natural language descriptions obtained via crowdsourcing. The motivation for this work is to develop a testbed for image sequence description systems, where the task is to generate natural language descriptions for animated GIFs or video clips. To ensure a high quality dataset, we developed a series of novel quality controls to validate free-form text input from crowdworkers. We show that there is unambiguous association between visual content and natural language descriptions in our dataset, making it an ideal benchmark for the visual content captioning task. We perform extensive statistical analyses to compare our dataset to existing image and video description datasets. Next, we provide baseline results on the animated GIF description task, using three representative techniques: nearest neighbor, statistical machine translation, and recurrent neural networks. Finally, we show that models fine-tuned from our animated GIF description dataset can be helpful for automatic movie description.