Ben Kröse

Sumit Mehra, Jantine van den Helder, Ben J. A. Kröse et al.

Sufficient physical activity can prolong the ability of older adults to live inde-pendently. Community-based exercise programs can be enhanced by regularly performing exercises at home. To support such a home-based exercise program, a blended intervention was developed that combined the use of a tablet application with a personal coach. The purpose of the current study was to explore to which extent tablet engagement predicted exercise adherence and physical activity. The results show that older adults (n=133; M=71 years of age) that participated 6 months in a randomized controlled trial, performed at average 12 home-based ex-ercised per week and exercised on average 3 days per week, thereby meeting WHO guidelines. They used the tablet app on average 7 times per week. Multiple linear regressions revealed that the use of the app statistically predicted the num-ber of exercises that were performed and the number of exercise days. Physical activity, however, did not increase and also could not be predicted by exercise frequency or app use. We conclude that engagement with a tablet can contribute to sustained exercise behavior.

Ninghang Hu, Gwenn Englebienne, Zhongyu Lou et al.

We present a novel hierarchical model for human activity recognition. In contrast to approaches that successively recognize actions and activities, our approach jointly models actions and activities in a unified framework, and their labels are simultaneously predicted. The model is embedded with a latent layer that is able to capture a richer class of contextual information in both state-state and observation-state pairs. Although loops are present in the model, the model has an overall linear-chain structure, where the exact inference is tractable. Therefore, the model is very efficient in both inference and learning. The parameters of the graphical model are learned with a Structured Support Vector Machine (Structured-SVM). A data-driven approach is used to initialize the latent variables; therefore, no manual labeling for the latent states is required. The experimental results from using two benchmark datasets show that our model outperforms the state-of-the-art approach, and our model is computationally more efficient.