A gesture recognition library for the therapy domain and its applications

Abstract

The computational implementation of human body gestures recognition has been a challenge for several years. Nowadays, thanks to the development of RGB-D cameras it is possible to acquire a set of data that represents a human position in space. Despite that, these cameras provide raw data, still being a problem to identify in real-time a specific pre-defined user movement continuously which can then be applied in applications as, for example, the tracking of physiotherapeutic movements or exercises. This work presents two new techniques to identify gestures, both having physiotherapeutic concerns about the performed exercise; one is based on physiotherapeutic standards, the biomechanical planes, while the other aims to recognize the functional exercises and is based on a concept called checkpoints. Both these techniques were tested and validated by physiotherapists from the Physiotherapy Department at the Federal University of Pernambuco. The techniques were also integrated in a library which was then used in two case studies and two general applications where their applicability was tested in physiotherapeutic and non-physiotherapeutic domains obtaining good results and showing that they can be used on general applications as well.