Handleap: towards contact-free gesture interaction with earphones via acoustic sensing

The lack of effective interaction methods restricts the direct operations of the earphones. Deploying acoustic-based interaction methods on earphones is a promising solution because of the ubiquitous, low-cost, and contact-free nature of the interaction. In this study, we propose HandLeap, a novel earphone interaction scheme to obtain personal classifiers using the gesture samples of a new user. We evaluated the efficacy and efficiency of HandLeap through two user experiments: evaluating the recognition accuracy in both the cold-start and warm-start modes and comparing them with the touch-on input schemes. HandLeap has been shown to help users access various functionalities efficiently in an earphone or comparable smart devices.