摘要 |
In a system utilising gesture recognition as a control mechanism, training of the system may be conducted whilst the system is in operation. The training method comprises obtaining a first frame 110 including an image of an object within a field of view; identifying the object as a hand by applying relevant algorithms 120, 130; storing image-related information of the identified hand 140; obtaining a second frame 150 including an image of an object within a field of view and identifying the object in the second frame as a hand 170 by using the stored information 160; and controlling the device 180 based on the hand identified in the frames. The values of a set of features of the objects may be compared to a threshold to make the identification. Machine learning techniques may be adopted as part of the method. The method ensures that the system is attuned to the userâ s specific environment and identifying features.
|