LSTM-Based Early Recognition of Motion Patterns

LSTM-Based Early Recognition of Motion Patterns
Markus Weber, Christopher Schölzel, Marcus Liwicki, Seiichi Uchida, Didier Stricker
Proceedings of the 22nd International Conference on Pattern Recognition International Conference on Pattern Recognition (ICPR-22), 22nd, August 24-28, Stockholm, Sweden

Abstract:

In this paper a method for early recognition of motion templates is presented. We define early recognition as an algorithm to provide recognition results before a motion sequence is completed. In our experiments we apply Long Short- Term Memory (LSTM) and optimize the training for the task of recognizing the motion template as early as possible. The evaluation has shown that the recognition accuracy for a frameby- frame classification the LSTM achieves a recognition accuracy of 88% if no training data of the person him/herself is included, and 92% if the training data also contains motion sequences of the person. Furthermore, the average earliness - the number of time steps it takes before the LSTM correctly classifies a motion pattern - is around 24.77 time steps, which is less than a second with the used tracking technology, i.e., the Microsoft Kinect.