ASL Alphabet gesture recognition on NZ ASL dataset
90AccuracyRF-JA+C(h-h)
Evaluation results
| Method | Links | |
|---|---|---|
| RF-JA+C(h-h)Hardware=color glove and depth-sensing Kinect camera, Training/validation split=50-502017.10 | 90 | |
| deepCNN2017.10 | 82.5 | |
| Stanford deepCNNBackground subtraction=false, Training dataset=ILSVRC20122017.10 | 72 | |
| RF-JA+C(l-o-o)Hardware=color glove and depth-sensing Kinect camera, Data protocol=omitted specific data2017.10 | 70 |