Abstract: In this research, we propose a hybrid CNN and LSTM model for word-level Ethiopian sign language recognition. The recognition system has four major components: preprocessing, feature ...