Generative Adversarial Network for Future Hand Segmentation from Egocentric Video

TL;DR

This paper presents EgoGAN, a deep generative model that predicts future hand masks from egocentric videos by modeling future head motions, outperforming previous methods on key datasets.

Contribution

Introduces EgoGAN, a novel deep generative model that anticipates future hand masks by incorporating stochastic head motion prediction in egocentric videos.

Findings

EgoGAN outperforms state-of-the-art methods in future hand mask prediction.

The model effectively captures the stochastic nature of head motions.

Detailed ablation studies validate the design choices of EgoGAN.

Abstract

We introduce the novel problem of anticipating a time series of future hand masks from egocentric video. A key challenge is to model the stochasticity of future head motions, which globally impact the head-worn camera video analysis. To this end, we propose a novel deep generative model -- EgoGAN, which uses a 3D Fully Convolutional Network to learn a spatio-temporal video representation for pixel-wise visual anticipation, generates future head motion using Generative Adversarial Network (GAN), and then predicts the future hand masks based on the video representation and the generated future head motion. We evaluate our method on both the EPIC-Kitchens and the EGTEA Gaze+ datasets. We conduct detailed ablation studies to validate the design choices of our approach. Furthermore, we compare our method with previous state-of-the-art methods on future image segmentation and show that our method can more accurately predict future hand masks.