SlowFast Rolling-Unrolling LSTMs for Action Anticipation in Egocentric Videos

Abstract

Action anticipation in egocentric videos is a difficult task due to the inherently multi-modal nature of human actions. Additionally, some actions happen faster or slower than others depending on the actor or surrounding context which could vary each time and lead to different predictions. Based on this idea, we build upon RULSTM architecture, which is specifically designed for anticipating human actions, and propose a novel attention-based technique to evaluate, simultaneously, slow and fast features extracted from three different modalities, namely RGB, optical flow, and extracted objects. Two branches process information at different time scales, i.e., frame-rates, and several fusion schemes are considered to improve prediction accuracy. We perform extensive experiments on EpicKitchens-55 and EGTEA Gaze+ datasets, and demonstrate that our technique systematically improves the results of RULSTM architecture for Top-5 accuracy metric at different anticipation times.

Accepted to EPIC@ICCV 2021

References in corpus (6)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Two-Stream Convolutional Networks for Action Recognition in Videos
• Rolling-Unrolling LSTMs for Action Anticipation from First-Person Video
• Next-Active-Object prediction from Egocentric Videos
• RED: Reinforced Encoder-Decoder Networks for Action Anticipation
• Visual Forecasting by Imitating Dynamics in Natural Sequences