Abstract. This paper introduces a new framework for image classification using local visual descriptors. The pipeline first performs a nonlinear feature transformation on descripto...
A new general theory about restoration of network paths is first introduced. The theory pertains to restoration of shortest paths in a network following failure, e.g., we prove th...
Clustering ensembles has been recently recognized as an emerging approach to provide more robust solutions to the data clustering problem. Current methods of clustering ensembles ...
This paper introduces a method to correct over-exposure in an existing photograph by recovering the color and lightness separately. First, the dynamic range of well exposed region...
This paper introduces an approach called Clustering and Co-evolution to Construct Neural Network Ensembles (CONE). This approach creates neural network ensembles in an innovative ...