Visual tracking is a critical task in many computer vision applications such as surveillance and robotics. However, although the robustness to local corruptions has been improved, prevailing trackers are still sensitive to large scale corruptions, such as occlusions and illumination variations. In this paper, we propose a novel robust object tracking technique depends on subspace learning-based appearance model. Our contributions are twofold. First, mask templates produced by frame difference are introduced into our template dictionary. Since the mask templates contain abundant structure information of corruptions, the model could encode information about the corruptions on the object more efficiently. Meanwhile, the robustness of the tracker is further enhanced by adopting system dynamic, which considers the moving tendency of the object. Second, we provide the theoretic guarantee that by adapting the modulated template dictionary system, our new sparse model can be solved by the accelerated proximal gradient algorithm as efficient as in traditional sparse tracking methods. Extensive experimental evaluations demonstrate that our method significantly outperforms 21 other cutting-edge algorithms in both speed and tracking accuracy, especially when there are challenges such as pose variation, occlusion, and illumination changes.
Prediction of physiological responses can have a number of applications in the health and medical fields. However, this can be a challenging task due to interdependencies between these responses, physical activities, environmental factors and the individual's mental state. In this work, we focus on forecasting physiological responses in dynamic scenarios where individuals are performing exercises and complex activities of daily life. We minimize the effect of environmental and physiological factors in order to focus on the effect of physical activities. In particular, we focus on forecasting heart rate and respiratory rate due to their relevance in medical and fitness training. We aim to forecast these physiological responses up to 60 s into the future, study the effect of different predictors that incorporate different sensing modalities and different amounts of historical data and analyze the performance of various strategies for prediction. Activity information is incorporated by clustering the data streams and fitting different predictive models per cluster. The effect of clustering is also studied by performing a hierarchical analysis on the clustering parameter, and we observe that activity clustering does improve the performance in our proposed methodology when predicting physiological response across modalities.
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.