Communicating physical gestures as compressed data streams

摘要

Computer-readable media and systems for communicating physical gestures as compressed data streams are provided. In particular, a technique is employed for generating replicated physical gestures on a user-interface display to visibly communicate sign language without transmitting information standard to typical video sessions, such as audio or video frames. In general, the technique entails the following steps: receiving coordinate representations of user-initiated physical gestures (e.g., hand movements); deciphering the received coordinate representations to identify corresponding motion patterns; and conveying the motion patterns to a presentation component for rendering. Rendering involves replicating the physical gestures by providing animations that carry out the motion patterns. The coordinate representations are generally received from a mobile device that captures physical gestures of a sender by employing a position-tracking interface. Upon capturing the physical gestures, the mobile device maps the coordinate representations to the physical gestures and records/transmits the coordinate representations at a high-speed frame rate.

主权项

1. One or more computer-readable media having computer-executable instructions embodied thereon that, when executed, perform a method for communicating physical gestures across a wireless network as a data stream, the method comprising:
receiving within the data stream coordinate representations of one or more user-initiated physical gestures captured by a position-tracking interface of a mobile device, wherein the physical gestures are captured by a process that includes at least temporarily, recording the coordinate representations at a high-speed frame rate that is greater than about 14 frames per second; determining that bandwidth is unavailable for the data stream of the coordinate representations such that the coordinate representations are converted into reduced bandwidth streaming data based in part on isolating the coordinate representations from corollary information; deciphering the coordinate representations to identify corresponding predefined motion patterns via a mapping procedure comprising:
(1) aggregating the coordinate representations that correspond to a particular one of the physical gestures; and(2) comparing, the aggregated coordinate representations against a library of predefined motion patterns to determine a match; upon identifying corresponding predefined motion patterns each being a graphical depiction of movement, conveying the predefined motion patterns different from a composite of the original image to a presentation component for rendering; and generating animations from the graphical depiction of movement of each of the predefined motion patterns that match the coordinate representations of the one or more physical gestures in replicating the one or more physical gestures.