This paper presents a novel method for illuminationinvariant and contrast preserving feature extraction, aimed at improving performance of tracking under complex light condition. ...
Network structure construction and global state maintenance are expensive in large-scale, dynamic peer-to-peer (p2p) networks. With inherent topology independence and low state mai...
Variable importance measures for random forests have been receiving increased attention as a means of variable selection in many classification tasks in bioinformatics and relate...
We consider the task of topology discovery of sparse random graphs using end-to-end random measurements (e.g., delay) between a subset of nodes, referred to as the participants. T...
Proofs and empirical evidence are presented which show that a subset of algorithms can have identical performance over a subset of functions, even when the subset of functions is ...