Sign language recognition from skeletal data using graph and recurrent neural networks

TL;DR

This paper introduces a novel graph and recurrent neural network approach for recognizing sign language gestures from skeletal pose data, demonstrating high accuracy on the AUTSL dataset and highlighting the effectiveness of pose-driven methods.

Contribution

It proposes a Graph-GRU temporal network that models spatial and temporal dependencies in skeleton data for sign language recognition, a novel integration for this task.

Findings

Achieved high accuracy on AUTSL dataset

Effective modeling of spatial and temporal dependencies

Scalable framework for sign language recognition

Abstract

This work presents an approach for recognizing isolated sign language gestures using skeleton-based pose data extracted from video sequences. A Graph-GRU temporal network is proposed to model both spatial and temporal dependencies between frames, enabling accurate classification. The model is trained and evaluated on the AUTSL (Ankara university Turkish sign language) dataset, achieving high accuracy. Experimental results demonstrate the effectiveness of integrating graph-based spatial representations with temporal modeling, providing a scalable framework for sign language recognition. The results of this approach highlight the potential of pose-driven methods for sign language understanding.