We examine the problem of acoustic emanations of printers. We present a novel attack that recovers what a dotmatrix printer processing English text is printing based on a record o...
This study focuses on the segmentation and characterization of oil slicks on the sea surface from synthetic aperture radar (SAR) observations. In fact, an increase in viscosity du...
We describe an acoustic chord transcription system that uses symbolic data to train hidden Markov models and gives best-of-class frame-level recognition results. We avoid the extre...
Several algorithms have been proposed to learn to rank entities modeled as feature vectors, based on relevance feedback. However, these algorithms do not model network connections...
Accurate unsupervised learning of phonemes of a language directly from speech is demonstrated via an algorithm for joint unsupervised learning of the topology and parameters of a ...