We describe a threshold-based local algorithm for image binarization. The main idea is to compute a transition energy using pixel value differences taken from a neighborhood aroun...
: This paper presents a feature selection technique based on distributional differences for efficient machine learning. Initial training data consists of data including many featur...
Background: There is an urgent need for new prognostic markers of breast cancer metastases to ensure that newly diagnosed patients receive appropriate therapy. Recent studies have...
Lei Xu, Aik Choon Tan, Raimond L. Winslow, Donald ...
Increasing values and spread in leakage current makes it impossible to distinguish between faulty and fault-free chips using single threshold method. Neighboring chips on a wafer ...
We present and evaluate a framework and tool for combining multiple program analyses which allows the dynamic (on-line) adjustment of the precision of each analysis depending on t...