Images captured in foggy weather conditions exhibit losses in quality which are dependent on distance. If the depth and atmospheric conditions are known, one can enhance the images...
We propose a novel boosting algorithm which improves on current algorithms for weighted voting classification by striking a better balance between classification accuracy and the ...
In this paper, we describe the design and implementation of two mechanisms for fault-tolerance and recovery for complex scientific workflows on computational grids. We present our ...
Current search engines can hardly cope adequately with fuzzy predicates defined by complex preferences. The biggest problem of search engines implemented with standard SQL is that...
Global shape information is an effective top-down complement
to bottom-up figure-ground segmentation as well
as a useful constraint to avoid drift during adaptive tracking.
We p...