This paper introduces an elderly exercise evaluation system. To determine the quality of a performed exercise, the authors propose a novel system to generate and use Spatio-Temporal Polychromatic Trajectory (STPT) images. Usually, the elder people need to perform some exercises or take physiotherapy in order to stay healthy both physically and mentally. It becomes difficult to evaluate the quality of their exercise routine without the aid of a trained physiotherapist. The system aims to overcome this problem by allowing elders to record their exercise videos using an easy-to-use Graphical User Interface and evaluate the results. A dataset of 109 subjects performing four types of shoulder exercises several times was created. The videos are labelled as correct or incorrect and an STPT image is generated from each video. Using our newly introduced method, the movement of the elder person is projected into an image which can be input to a Convolutional Neural Network (CNN). The dataset is further augmented to increase accuracy. Using our proposed method, the best model achieved an F1 Score over 90% in three of the four exercises. The CNN is trained based on these clips and the models are added to the backend of the interface. The proposed system requires only an ordinary camera and a computer with an entry level GPU allowing it to be deployed at a large scale.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.