Fig 1.
Internal structure of the LTSM unit.
Fig 2.
Schematic diagram of the BiLTSM.
Fig 3.
Dynamic gesture recognition pipeline based on A-CBLN.
Table 1.
Specific parameter settings for each layer in A-CBLN.
Table 2.
Key parameters of the data glove used in this paper.
Table 3.
Detailed description of the seven dynamic gestures.
Fig 4.
Flow chart of the gesture data acquisition.
Fig 5.
Variation process of validation accuracy of different models.
Table 4.
Prediction results of different models on the test dataset.
Fig 6.
Effect of different convolutional kernel size on the A-CBLN algorithm.
Fig 7.
Effect of different number of neurons in BiLSTM on A-CBLN algorithm.
Fig 8.
Confusion matrix of A-CBLN on the test dataset.