Respiratory Motion Forecasting on RPM data Varian
0.139RMSELin [23]
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Lin [23]Network=3-layer LSTM, Sampling rate=30 Hz, Amount of data=1703 records, Signal amplitude=Rescaling between -1 and 1, Response time=280ms / 500ms2021.06 | 0.139 | 0.112 | 1.811 | — | |
| Sun [46]Network=1-layer MLP, Training method=Levenberg-Marq. & adapt. boosting, Sampling rate=30 Hz, Amount of data=data from 138 scans, Signal amplitude=Rescaling between -1 and 1, Response time=500ms2021.06 | 0.17 | — | 0.65 | 0.28 | |
| Lee [19]Network=LSTM-FCL, Training method=BPTT, Sampling rate=30 Hz, Amount of data=550 records, Signal amplitude=11.9mm to 25.9mm, Response time=210ms2021.06 | 0.28 | — | — | — |