A neural network based on SPD manifold learning for skeleton-based hand gesture recognition - Normandie Université Access content directly
Conference Papers Year : 2019

A neural network based on SPD manifold learning for skeleton-based hand gesture recognition

Abstract

This paper proposes a new neural network based on SPD manifold learning for skeleton-based hand gesture recognition. Given the stream of hand's joint positions, our approach combines two aggregation processes on respectively spatial and temporal domains. The pipeline of our network architecture consists in three main stages. The first stage is based on a convolutional layer to increase the discrim-inative power of learned features. The second stage relies on different architectures for spatial and temporal Gaussian aggregation of joint features. The third stage learns a final SPD matrix from skeletal data. A new type of layer is proposed for the third stage, based on a variant of stochastic gradient descent on Stiefel manifolds. The proposed network is validated on two challenging datasets and shows state-of-the-art accuracies on both datasets.
Fichier principal
Vignette du fichier
Nguyen_A_Neural_Network_Based_on_SPD_Manifold_Learning_for_Skeleton-Based_CVPR_2019_paper.pdf (557.02 Ko) Télécharger le fichier
Origin Files produced by the author(s)
Loading...

Dates and versions

hal-02456437 , version 1 (27-01-2020)

Identifiers

  • HAL Id : hal-02456437 , version 1

Cite

Xuan Son Nguyen, Luc Brun, Olivier Lezoray, Sébastien Bougleux. A neural network based on SPD manifold learning for skeleton-based hand gesture recognition. IEEE Conference on Computer Vision and Pattern Regognition (CVPR), 2019, Long Beach, United States. ⟨hal-02456437⟩
75 View
164 Download

Share

Gmail Mastodon Facebook X LinkedIn More