Real Time Classification
Our Group develops algorithms for processing sensor data on highly resource constrained sensor nodes, in order to provide only useful data back to the cloud, and save on communication costs. Our work on real-time classification has covered species detection and classification through bioacoustics signals, activity detection of livestock, and the use of compressive sensing and sparse representation methods for efficient classification.
N. Kottege, R. Jurdak, F. Kroon, D. Jones, “Automated detection of broadband clicks of freshwater fish using spectro-temporal features,” Journal of the Acoustical Society of America (JASA), 137, 2502 May 2015.
N. Kottege, R. Jurdak, F. Kroon, D. Jones, “Classification of Underwater Broadband Acoustic Sounds using Spectro-temporal Features,” In proceedings of the Seventh ACM International Conference on Underwater Networks and Systems (WUWNet), Los Angeles CA, November 2012.