Research on Optimizer Algorithm of Sign Language Recognition Model

Abstract

One of the most well-known tools for sign language recognition is neural network. There are many optimizer algorithms to improve the performance of the network. But previous studies of optimizer algorithms have not dealt with its convergence stability for sign language recognition. Therefore, we propose an optimizer algorithm for sign language recognition model, which has a good optimization effect on sign language recognition model based on LSTM neural network. We designed a sign language recognition model based on LSTM neural network, and carried out the experiment in the open dataset of SLR of University of science and technology of China. In the experiment, we added the classical optimizer algorithm for comparison. The experimental results show that. This paper proposes that the optimizer algorithm has better convergence stability than the classical algorithm, and has good adaptability to different input data.