Skip to main content
Fig. 4 | Journal of NeuroEngineering and Rehabilitation

Fig. 4

From: User activity recognition system to improve the performance of environmental control interfaces: a pilot study with patients

Fig. 4

Environment control interface performance. The users performed two different trials with the same goal: complete as many tasks as possible in 10 min by using the Manual mode and the AIDE mode (a). Both modes have been used to train a SVM model with Gaussian kernel and find the boundary between them [yellow line in (a)]. The interaction with the ECI was measured by steps that a user had to perform in order to complete the tasks (c). Furthermore, in AIDE mode the user’s intention prediction was performed through CRF model and, therefore, the confusion matrix of the model is obtained (b)

Back to article page