A very efficient strategy for very large data sets clustering: application to image segmentation
Michel M. Bruynooghe · 2003
The clustering strategy for very large data sets that is presented overcomes the limits of the classical strategy based on the clustering of the so-called strong patterns. The strategy presented leads to results of high statistical quality and is at least 25 times faster than the classical strategy in the case of the clustering of about 10000 pixels coming from a sensitometry image.>