Temporal Action Detection with Structured Segment Networks

TLDR

Detecting actions in untrimmed videos is an important yet challenging task. The paper proposes the Structured Segment Network (SSN), a framework that models action temporal structure with a structured temporal pyramid and introduces the Temporal Actionness Grouping (TAG) proposal scheme. SSN employs a structured temporal pyramid and a decomposed discriminative model with separate classifiers for action classification and completeness, integrated into a unified network trained end‑to‑end. The framework accurately distinguishes positive proposals from background or incomplete ones, achieving superior recognition and localization, and outperforms state‑of‑the‑art methods on THUMOS14 and ActivityNet.

Abstract

Detecting actions in untrimmed videos is an important yet challenging task. In this paper, we present the structured segment network (SSN), a novel framework which models the temporal structure of each action instance via a structured temporal pyramid. On top of the pyramid, we further introduce a decomposed discriminative model comprising two classifiers, respectively for classifying actions and determining completeness. This allows the framework to effectively distinguish positive proposals from background or incomplete ones, thus leading to both accurate recognition and localization. These components are integrated into a unified network that can be efficiently trained in an end-to-end fashion. Additionally, a simple yet effective temporal action proposal scheme, dubbed temporal actionness grouping (TAG) is devised to generate high quality action proposals. On two challenging benchmarks, THUMOS14 and ActivityNet, our method remarkably outperforms previous state-of-the-art methods, demonstrating superior accuracy and strong adaptivity in handling actions with various temporal structures.