Determination of run-length smoothing values for document segmentation
N. Papamarkos, J. Tzortzakis, Basilis Gatos · 2002
The segmentation is a pre-processing procedure in document analysis systems. In this paper, a document segmentation method is proposed which is based on run-length smoothing algorithm. Run-length based segmentation methods use heuristic techniques to determine the values of the smoothing variables. In this paper, we present an unsupervised technique, based on the distributions of the black and white run-lengths, for the calculation of proper values of the smoothing variables.