Duodepth: Static Gesture Recognition Via Dual Depth Sensors

摘要: Static gesture recognition is an effective non-verbal communication channel between a user and their devices; however many modern methods are sensitive to the relative pose of user’s hands with respect capture device, as parts can become occluded. We present two methodologies for via synchronized recording from depth cameras alleviate this occlusion problem. One more classic approach using iterative closest point registration accurately fuse clouds single PointNet architecture classification, other dual Point-Net classification without registration. On manually collected data-set 20,100 we show 39.2% reduction in misclassification fused cloud method, 53.4% PointNet, when compared standard camera pipeline.