This paper proposes three novel training methods, two of them based on the back-propagation approach and a third one based on information theory for Multilayer Perceptron (MLP) bin...
Psychologists have proposed that many human-object interaction activities form unique classes of scenes. Recognizing these scenes is important for many social functions. To enable...
Many data mining applications involve the task of building a model for predictive classification. The goal of such a model is to classify examples (records or data instances) into...
Elon S. Correa, Alex Alves Freitas, Colin G. Johns...
Support Vector Machines (SVMs) have been applied to solve the classification of volatile organic compounds (VOC) data in some recent studies. SVMs provide good generalization perfo...
Software available at http://disi.unitn.it/~uijlings or http://koen.me/research/
For object recognition, the current state-of-the-art is based on exhaustive search. However, to ...