How Much Temporal Long-Term Context is Needed for Action Segmentation?

TL;DR

This paper investigates the amount of long-term temporal context needed for effective action segmentation in videos, proposing a sparse attention transformer model that captures full video context and outperforms existing methods.

Contribution

Introduces a transformer-based model with sparse attention to efficiently capture full video context for action segmentation, demonstrating its superiority over local-window approaches.

Findings

Full context modeling improves segmentation accuracy.

Sparse attention effectively captures long-term dependencies.

Model outperforms state-of-the-art on three datasets.

Abstract

Modeling long-term context in videos is crucial for many fine-grained tasks including temporal action segmentation. An interesting question that is still open is how much long-term temporal context is needed for optimal performance. While transformers can model the long-term context of a video, this becomes computationally prohibitive for long videos. Recent works on temporal action segmentation thus combine temporal convolutional networks with self-attentions that are computed only for a local temporal window. While these approaches show good results, their performance is limited by their inability to capture the full context of a video. In this work, we try to answer how much long-term temporal context is required for temporal action segmentation by introducing a transformer-based model that leverages sparse attention to capture the full context of a video. We compare our model with the current state of the art on three datasets for temporal action segmentation, namely 50Salads, Breakfast, and Assembly101. Our experiments show that modeling the full context of a video is necessary to obtain the best performance for temporal action segmentation.