CFAD: Coarse-to-Fine Action Detector for Spatiotemporal Action Localization

TL;DR

CFAD is an end-to-end framework for spatio-temporal action localization that first estimates coarse action tubes and then refines them, achieving high accuracy and faster inference.

Contribution

Introduces a novel coarse-to-fine paradigm with two modules for efficient and accurate spatio-temporal action localization.

Findings

Achieves competitive results on UCF101-24, UCFSports, JHMDB-21 datasets.

Runs 3.3x faster than nearest competitors.

Effectively refines action tubes using key timestamp guidance.

Abstract

Most current pipelines for spatio-temporal action localization connect frame-wise or clip-wise detection results to generate action proposals, where only local information is exploited and the efficiency is hindered by dense per-frame localization. In this paper, we propose Coarse-to-Fine Action Detector (CFAD),an original end-to-end trainable framework for efficient spatio-temporal action localization. The CFAD introduces a new paradigm that first estimates coarse spatio-temporal action tubes from video streams, and then refines the tubes' location based on key timestamps. This concept is implemented by two key components, the Coarse and Refine Modules in our framework. The parameterized modeling of long temporal information in the Coarse Module helps obtain accurate initial tube estimation, while the Refine Module selectively adjusts the tube location under the guidance of key timestamps. Against other methods, theproposed CFAD achieves competitive results on action detection benchmarks of UCF101-24, UCFSports and JHMDB-21 with inference speed that is 3.3x faster than the nearest competitors.