Abstract
Human activity recognition is an extensively researched topic in the last decade. Recent methods employ supervised and unsupervised deep learning techniques in which spatial and temporal dependency is modeled. This paper proposes a novel approach for human activity recognition using skeleton data. The method combines supervised and unsupervised learning algorithms in order to provide qualitative results and performance in real time. The proposed method involves a two-stage framework: the first stage applies an unsupervised clustering technique to group up activities based on their similarity, while the second stage classifies data assigned to each group using graph convolutional networks. Different clustering techniques and data augmentation strategies are explored for improving the training process. The results were compared against the state of the art methods and the proposed model achieved 90.22% Top-1 accuracy performance for NTU-RGB+D dataset (the performance was increased by approximately 9% compared with the baseline graph convolutional method). Moreover, inference time and total number of parameters stay within the same magnitude order. Extending the initial set of activities with additional classes is fast and robust, since there is no required retraining of the entire architecture but only to retrain the cluster to which the activity is assigned.
Highlights
Some of the activities are misclassified due to their similar behavior of the subject movements. The problem, in this case, is that the skeleton representation is polluted with noise due to the inaccurate detection of the joints
The t-distributed stochastic neighbor embedding’ [26] (t-SNE) algorithm begins by computing a probability distribution that resembles the similarity of points
The similarity between two points xi and x j is equal to the probability P(i|j), meaning that the point xi would pick x j as its neighbour, if neighbors are picked to fit a Gaussian centered at xi . t-SNE will compute these probabilities in the high dimensional space, the original one, and it will compute the same probabilities of the projected points in the 2D space where we want to visualize the data
Summary
The costs of social care and hospitalization constantly increase. In order to reduce these costs, there is a need to maintain elderly people living in their homes for a long time. In order to make this possible, intensive research is conducted in the interest of creating assistive systems that are able to perform continuous monitoring of health and activity performed by elderly at home and to detect early stages of abnormal situations. Based on a study of the World Health Organisation, assistive systems enable people to live healthily and independently, reducing the need for formal health and support services [1]. Different systems were already proposed using robotic platforms:
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
