Table 8 Comparison of the models.

From: A deep learning framework for Ethiopian sign language recognition using skeleton-based representation

Models

Comparison metrics

Training

accuracy (%)

Testing

accuracy (%)

Average

training

duration

Number of

total

parameters

LSTM-based

EthSLR model

98

91

2 h

2,352,836

BiLSTM-based

EthSLR model

97

92

2 h

3,904,148

CNN-LSTM-based

EthSLR model

99

93

2 h

4,291,156

GRU-based EthSLR

model

99

94

3 h

4,364,068

Back to article page