System theoretic approaches to action recognition model the dynamics of a scene with linear dynamical systems (LDSs) and perform classification using metrics on the space of LDSs, ...
Rizwan Chaudhry, Avinash Ravichandran, Gregory D. ...
Most modern computer vision systems for high-level
tasks, such as image classification, object recognition and
segmentation, are based on learning algorithms that are
able to se...
Now the classification of different tumor types is of great importance in cancer diagnosis and drug discovery. It is more desirable to create an optimal ensemble for data analysis ...
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
Anticipating the availability of large questionanswer datasets, we propose a principled, datadriven Instance-Based approach to Question Answering. Most question answering systems ...