Fine-grained Gesture Recognition Using WiFi

Gesture recognition has drawn increasingly attention in human-computer interaction (HCI) and can support a variety of emerging applications such as smart horne and mobile gaming. Traditional approaches usually involve wearable sensors and specialized hardware installations. This paper presents fine-grained finger gesture recognition by using a single commodity WiFi device without requiring user to wear any sensor. Our low-cost system, WiFinger, takes advantages of the detailed channel state information (CSI) available from commodity WiFi devices and the prevalence of WiFi infrastructure. It senses and identifies subtle movements of finger gestures by examining the unique patterns exhibited in the detailed CSI. We devise environmental noise removal mechanism to mitigate the efTect of signal dynamic due to the environment changes. We also design algorithm to capture the intrinsic gesture behavior to deal with individual diversity and gesture inconsistency. Our experimental evaluation in a horne environment demonstrates that our system can achieve over 93% recognition accuracy and is robust to both environment changes and individual diversity.