Unsupervised Learning from Narrated Instruction Videos

TL;DR

This paper presents an unsupervised method that learns task steps from narrated instruction videos by jointly clustering video and narration data, validated on a new large-scale dataset.

Contribution

It introduces a novel unsupervised approach leveraging both video and narration, and provides a new challenging dataset for instruction video analysis.

Findings

Successfully discovers main task steps without supervision

Effectively locates steps within input videos

Demonstrates robustness on complex real-world data

Abstract

We address the problem of automatically learning the main steps to complete a certain task, such as changing a car tire, from a set of narrated instruction videos. The contributions of this paper are three-fold. First, we develop a new unsupervised learning approach that takes advantage of the complementary nature of the input video and the associated narration. The method solves two clustering problems, one in text and one in video, applied one after each other and linked by joint constraints to obtain a single coherent sequence of steps in both modalities. Second, we collect and annotate a new challenging dataset of real-world instruction videos from the Internet. The dataset contains about 800,000 frames for five different tasks that include complex interactions between people and objects, and are captured in a variety of indoor and outdoor settings. Third, we experimentally demonstrate that the proposed method can automatically discover, in an unsupervised manner, the main steps to achieve the task and locate the steps in the input videos.