A recurrent neural network model for sign language classification

Abstract

Sign language classification is the most critical task of sign language information recognition. This task involves the extraction of information features of sign language materials. It mainly focuses on describing the underlying features of sign language and improving the accuracy of sign language classification. The sign language classification model based on recurrent neural network has the problems of complex model input data processing, complex network model and slow training speed. This paper presents a classification model based on recurrent neural network bidirectional long-term and short-term memory network and puts forward an optimized scheme from network data processing, network structure and network training method. Experiments show the effectiveness of the proposed model.