An Interactive Care System Based on a Depth Image and EEG for Aged Patients with Dementia

Due to the limitations of the body movement and functional decline of the aged with dementia, they can hardly make an efficient communication with nurses by language and gesture language like a normal person. In order to improve the efficiency in the healthcare communication, an intelligent interactive care system is proposed in this paper based on a multimodal deep neural network (DNN). The input vector of the DNN includes motion and mental features and was extracted from a depth image and electroencephalogram that were acquired by Kinect and OpenBCI, respectively. Experimental results show that the proposed algorithm simplified the process of the recognition and achieved 96.5% and 96.4%, respectively, for the shuffled dataset and 90.9% and 92.6%, respectively, for the continuous dataset in terms of accuracy and recall rate.