Skip to content

Latest commit

 

History

History
20 lines (16 loc) · 741 Bytes

README.md

File metadata and controls

20 lines (16 loc) · 741 Bytes

Decoupled and Boosted Learning for Skeleton-based Dynamic Hand Gesture Recognition

Introduction

In this paper, we propose a lightweight dual-stream framework based on decoupled and boosted learning for skeleton-based dynamic hand gesture recognition. We evaluate our model on three challenging datasets: SHREC’17 Track dataset, FPHA dataset, and DHG-14/28 dataset. Experimental results show the superiority of our method.

Requirements

  • Python 3.8
  • Tensorflow 2.4.1
  • numpy
  • tqdm
  • scipy
  • opencv

Test Model

  • You can directly download the trained models from here.
  1. Run the following command to test models.
python test.py