In developing automated systems to recognize the emotional content of music, we are faced with a problem spanning two disparate domains: the space of human emotions and the acoust...
Erik M. Schmidt, Douglas Turnbull, Youngmoo E. Kim
In traditional peer-to-peer search networks, operations focus on properly labeled files such as music or video, and the actual search is often limited to text tags. The explosive ...
Recently, boosting has come to be used widely in object-detection applications because of its impressive performance in both speed and accuracy. However, learning weak classifier...
: Problem statement: For a sensor network comprising autonomous and self-organizing data sources, efficient similarity-based search for semantic-rich resources (such as video data)...
This paper describes the baselines proposed for the ResPubliQA 2009 task. These baselines are purely based on information retrieval techniques. The selection of an adequate retrie...