Human posture tracking with flexible sensors for motion recognition

The integration of conventional clothes with flexible electronics is a promising solution as future?generation computing platform. However, the problem user authentication on this novel platform still underexplored. This work uses sensors to track human posture and achieves goal authentication. We capture movement pattern by four stretch around shoulder one elbow. introduce long short?term memory fully convolutional network (LSTM?FCN), which directly takes noisy sparse sensor data input verifies its consistency user's predefined patterns. method can identify matching patterns even if there are large intrapersonal variations. accuracy LSTM?FCN reaches 98.0%, 10.7% 6.5% higher than that dynamic time warping dependent.