作者: Tracy L. Westeyn , Gregory D. Abowd , Thad E. Starner , Jeremy M. Johnson , Peter W. Presti
DOI: 10.1007/S00779-011-0386-0
关键词:
摘要: Previous research has established the connection between way in which children interact with objects and potential early identification of autism. Those findings motivate our own work to develop "smart toys," embedded wireless sensors that are safe enjoyable for very small children, allow detailed interaction data be easily recorded. These sensor-enabled toys provide opportunities autism by reducing effort required collect analyze a child's interactions objects. In future, such may useful part clinical in-home assessment tools. this paper, we discuss design collection smart can used automatically characterize child is playing. We use statistical models objective, quantitative measures object play interactions. also developed tool view rich forms annotated later analysis. report results recognition experiments on more than fifty sessions conducted adults as well using approach support video annotation other applications.