ORIGINAL RESEARCH published: 12 February 2019 doi: 10.3389/fnins.2019.00073 FLGR: Fixed Length Gists Representation Learning for RNN-HMM Hybrid-Based Neuromorphic Continuous Gesture Recognition Guang Chen 1,2*†, Jieneng Chen 3†, Marten Lienen 2†, Jörg Conradt 4, Florian Röhrbein 2 and Alois C. Knoll 2* 1 College of Automotive Engineering, Tongji University, Shanghai, China, 2 Chair of Robotics, Artificial Intelligence and Real-time Systems, Technische Universität München, Munich, Germany, 3 College of Electronics and Information Engineering, Tongji University, Shanghai, China, 4 Department of Computational Science and Technology, KTH Royal Institute of Technology, Stockholm, Sweden Edited by: Runchun Mark Wang, A neuromorphic vision sensors is a novel passive sensing modality and frameless sensors Western Sydney University, Australia with several advantages over conventional cameras. Frame-based cameras have an Reviewed by: average frame-rate of 30 fps, causing motion blur when capturing fast motion, e.g., hand Hesham Mostafa, gesture. Rather than wastefully sending entire images at a fixed frame rate, neuromorphic University of California, San Diego, United States vision sensors only transmit the local pixel-level changes induced by the movement in Arren Glover, a scene when they occur. This leads to advantageous characteristics, including low Fondazione Istituto Italiano di Tecnologia, Italy energy consumption, high dynamic range, a sparse event stream and low response *Correspondence: latency. In this study, a novel representation learning method was proposed: Fixed Length Guang Chen Gists Representation (FLGR) learning for event-based gesture recognition. Previous
[email protected] methods accumulate events into video frames in a time duration (e.g., 30 ms) to make Alois C.