Application of direction constrained and bipolar waves for pattern recognition
István Petrás, T. Roska · 2002
Direction constrained and bipolar waves are introduced. Their possible applications for direction selective curvature and concavity detection as well as region segmentation are shown. A cellular neural (CNN) algorithm frame for feature-based object decomposition is presented. Algorithms are tested on the 64/spl times/64 CNNUM (CNN Universal Machine) chip.