TAO

Translating fine-grained activity detection (e.g., phone ring, talking interspersed with silence and walking) into semantically meaningful richer contextual information on a call for 20 minutes while exercising) is essential towards enabling range of healthcare human-computer interaction applications. Prior work has proposed building ontologies or temporal analysis patterns limited success in capturing complex real-world context patterns. We present TAO, hybrid system that leverages OWL-based clustering approaches to detect high-level contexts from human activities. TAO can characterize sequential activities happen one after the other are interleaved occur parallel set more accurately than prior work. evaluate datasets (Casas Extrasensory) show our achieves, average, 87% 80% accuracy detection, respectively. deploy setting eight participants using three hours each, demonstrating TAO's ability capture real world. Finally, showcase usefulness contexts, we prototype wellness applications assess productivity stress metrics calculated provided by much closer ground truth (on average within 1.1%), as compared baseline approach 30%).