Unsupervised clustering can be significantly improved using supervision in the form of pairwise constraints, i.e., pairs of instances labeled as belonging to same or different clu...
In this paper, we systematically study the effect of poorly registered faces on the training and inferring stages of traditional face recognition algorithms. We then propose a nov...
Abstract. The Unpredictability Measure computation algorithm applied to psychoacoustic model-based broadband noise attenuation is discussed. A learning decision algorithm based on ...
Andrzej Czyzewski, Marek Dziubinski, Lukasz Litwic...
Graph matching is an important problem in computer
vision. It is used in 2D and 3D object matching and recognition.
Despite its importance, there is little literature on
learnin...
We present an approach for the supervised online learning of object representations based on a biologically motivated architecture of visual processing. We use the output of a rece...