Actor and Action Video Segmentation from a Sentence

TL;DR

This paper introduces a novel method for pixel-level actor and action segmentation in videos guided by natural language sentences, enabling more flexible and fine-grained segmentation beyond fixed vocabularies.

Contribution

It presents a fully-convolutional encoder-decoder model that infers segmentation from sentences and extends datasets with natural language descriptions for evaluation.

Findings

Model achieves high-quality sentence-guided segmentations.

Demonstrates strong generalization to unseen actors and actions.

Outperforms state-of-the-art in traditional segmentation tasks.

Abstract

This paper strives for pixel-level segmentation of actors and their actions in video content. Different from existing works, which all learn to segment from a fixed vocabulary of actor and action pairs, we infer the segmentation from a natural language input sentence. This allows to distinguish between fine-grained actors in the same super-category, identify actor and action instances, and segment pairs that are outside of the actor and action vocabulary. We propose a fully-convolutional model for pixel-level actor and action segmentation using an encoder-decoder architecture optimized for video. To show the potential of actor and action video segmentation from a sentence, we extend two popular actor and action datasets with more than 7,500 natural language descriptions. Experiments demonstrate the quality of the sentence-guided segmentations, the generalization ability of our model, and its advantage for traditional actor and action segmentation compared to the state-of-the-art.