FMHash: Deep Hashing of In-Air-Handwriting for User Identification.

Gesture biometrics are gaining popularity with gesture input interface on mobile and Virtual Reality (VR) platforms that lack a keyboard or touchscreen to type a password for user authentication. However, less attention is paid to the gesture-based user identification problem, which essentially requires indexing and searching the gesture motion templates in a large database efficiently. In this paper, we propose FMHash, a user identification framework that can generate a compact binary hash code from a piece of in-air-handwriting of an ID string, which allows fast search in a database of in-air-handwriting templates through a hash table. To demonstrate the effectiveness of the framework, we implemented a prototype and report preliminary results (~98% precision and ~93% recall). More detailed evaluation, comparison, and improvement is working-in-progress. The ability of hashing in-air-handwriting pattern to binary code can be used to achieve convenient sign-in and sign-up with in-air-handwriting gesture ID on future mobile and VR devices.