``Listenin'' to domestic enviroments from remote locations
MetadataShow full item record
This paper describes ListenIn, work in progress using audio as a monitoring medium, with emphasis on domestic environments inhabited by elder parents. The primary goal of this monitoring is to provide a continuous but peripheral awareness of the monitored site and a remote location, or to a mobile user. Sound gathering and classification in the home is done in a distributed architecture server with multiple components. At transitions of activity, as measured by change in sound, a remote server receives and plays a few seconds of an ``iconic'' sound, the actual sound, or a ``garbled'' version of the actual sound, depending on confidence in the classification and whether speech is detected.