Figures & data
Figure 1. Stacked autoencoder architecture.
Figure 2. Potential sabotage motions.
Figure 3. Real-time hand tracking results.
Figure 4. Conversion of captured hand motion data to a time-series format training set.
Figure 5. Proposed SAE model architecture.
Table 1. Four proposed network architectures
Figure 6. Motion recognition output. Different colored lines represent the probability of different motions (red line: probability of pushing motion; blue line: probability of grasping motion; green line: probability of cutting motion; yellow line: probability of patting motion; pink line: probability of turning motion).
Table 2. Motion recognition accuracy
Figure 7. Verification dataset generated for nonconsecutive frame loss.
Figure 8. Verification dataset generated for consecutive frame loss.
Table 3. Verification results for nonconsecutive frame loss
Table 4. Verification results for consecutive frame loss