Long Short-Term Transformer for Online Action Detection
Published in
Neural Information Processing Systems(2021)External Links:
TL
TL;DR
LSTR provides an effective and efficient method to model long videos with fewer heuristics, which is validated by extensive empirical analysis and achieves state-of-the-art performance on three standard online action detection benchmarks.
Abstract
We present Long Short-term TRansformer (LSTR), a temporal modeling algorithm for online action detection, which employs a long- and short-term memory mechanism to model prolonged sequence data. It consists of an LSTR encoder that dynamically leverages coarse-scale historical information from an extended temporal window (e.g., 2048 frames spanning of up to 8 minutes), together with an LSTR decoder that focuses on a short time window (e.g., 32 frames spanning 8 seconds) to model the fine-scale characteristics of the data. Compared to prior work, LSTR provides an effective and efficient method to model long videos with fewer heuristics, which is validated by extensive empirical analysis. LSTR achieves state-of-the-art performance on three standard online action detection benchmarks, THUMOS'14, TVSeries, and HACS Segment. Code has been made available at: https://xumingze0308.github.io/projects/lstr
References75 items
1
Towards Long-Form Video Understanding
2
VidTr: Video Transformer Without Convolutions
3
Temporally smooth online action detection using cycle-consistent future anticipation
4
ViViT: A Video Vision Transformer
5