This paper presents a novel approach to pedestrian classification which involves utilizing the synthesized virtual samples of a learned generative model to enhance the classificat...
Background: Recursive Feature Elimination is a common and well-studied method for reducing the number of attributes used for further analysis or development of prediction models. ...
The construction of a text classifier usually involves (i) a phase of term selection, in which the most relevant terms for the classification task are identified, (ii) a phase ...
Background: In class prediction problems using microarray data, gene selection is essential to improve the prediction accuracy and to identify potential marker genes for a disease...
Many problems in vision involve the prediction of a class label for each frame in an unsegmented sequence. In this paper, we develop a discriminative framework for simultaneous se...