Gesture recognition using recurrent neural networks

TLDR

Gesture recognition is more challenging than posture recognition because it must process dynamic movements. The paper presents a gesture recognition method for Japanese sign language capable of recognizing continuous gestures. The authors build a neural‑network posture recognizer for 42 finger‑alphabet symbols, extend it to a gesture recognizer mapping gestures to words, and employ a recurrent neural network to handle dynamic processes for continuous gesture recognition. The results show that the proposed system can recognize continuous gestures.

Abstract

A gesture recognition method for Japanese sign language is presented. We have developed a posture recognition system using neural networks which could recognize a finger alphabet of 42 symbols. We then developed a gesture recognition system where each gesture specifies a word. Gesture recognition is more difficult than posture recognition because it has to handle dynamic processes. To deal with dynamic processes we use a recurrent neural network. Here, we describe a gesture recognition method which can recognize continuous gesture. We then discuss the results of our research.