Table 1.
Literature comparison table.
Fig 1.
Map of selected datasets from ASL [26].
Fig 2.
Self-constructed partial dataset maps.
Fig 3.
Sample diagram of a portion of the NUS-II dataset [27].
Fig 4.
Gesture recognition flowchart.
Fig 5.
Multiscale convolutional framework.
Fig 6.
Flowchart of gesture skin color segmentation.
Fig 7.
Standard convolutional map.
Fig 8.
Deeply separable convolutional maps.
Fig 9.
Multiscale convolution module.
Fig 10.
Comparison of ReLU and ELU activation functions.
Fig 11.
MS-MobileNet structure diagram.
Fig 12.
Skin color extraction process result chart.
Fig 13.
Skin color extraction performance comparison chart.
Table 2.
Performance comparison on NUS-II and creative Senz3D datasets.
Table 3.
Performance comparison of different models on ASL-M dataset.
Table 4.
Comparison of recognition efficiency of models.
Fig 14.
Confusion matrix for letter gestures [30].
Table 5.
Alphabet gesture recognition performance metrics for different categories.
Table 6.
Ablation experiment.