OadTR: Online Action Detection with Transformers

TL;DR

OadTR introduces a Transformer-based encoder-decoder framework for online action detection, effectively capturing historical and future context to improve accuracy and speed over RNN-based methods.

Contribution

The paper presents a novel Transformer-based architecture for online action detection, addressing RNN limitations and enhancing performance on challenging datasets.

Findings

Outperforms state-of-the-art methods in mAP and mcAP

Achieves higher training and inference speeds

Effectively models historical and future context

Abstract

Most recent approaches for online action detection tend to apply Recurrent Neural Network (RNN) to capture long-range temporal structure. However, RNN suffers from non-parallelism and gradient vanishing, hence it is hard to be optimized. In this paper, we propose a new encoder-decoder framework based on Transformers, named OadTR, to tackle these problems. The encoder attached with a task token aims to capture the relationships and global interactions between historical observations. The decoder extracts auxiliary information by aggregating anticipated future clip representations. Therefore, OadTR can recognize current actions by encoding historical information and predicting future context simultaneously. We extensively evaluate the proposed OadTR on three challenging datasets: HDD, TVSeries, and THUMOS14. The experimental results show that OadTR achieves higher training and inference speeds than current RNN based approaches, and significantly outperforms the state-of-the-art methods in terms of both mAP and mcAP. Code is available at https://github.com/wangxiang1230/OadTR.