Portable Gesture Recognition Based on Flexible Stretchable Electronic Skin

In response to the large size and high price of traditional wearable gesture recognition systems, we presented a gesture recognition system based on self-developed flexible electronic skin (e-skin) with the advantages of easy portability, low cost and high recognition rate. Fourteen flexible sensors were integrated into the e-skin gesture recognition glove, which could monitor the movement of each finger joint. Meanwhile, an accelerometer was integrated, which could collect rollover information of the wrist. Furthermore, we innovatively proposed an improved dynamic time warping (DTW) algorithm to build a gesture template library, which is used to recognize complex and continuously changing gestures. Finally, the recognition results are displayed in the form of text, pictures and even the voice broadcast. Continuous recognition experiments on eight daily gestures were carried out, and the segmentation accuracy and recognition accuracy of continuous gestures is 97.49% and 93.73% respectively, which can meet the needs of deaf-mute people for normal communication with the outside world.