$Id: README,v 1.1 2000/09/19 22:22:05 maosong Exp maosong $ Package: PSET version 1.01 Author: Song Mao (kanungo@cfar.umd.edu) Organization: University of Maryland, College Park, MD Web: http://www.cfar.umd.edu/~kanungo Date: 18 September, 2000 PSET is a software package for evaluating page segmentation algorithms. It has two major functions: automatically training page segmentation algorithms on a given training dataset, and testing page segmentation algorithms on a given test dataset. Three research page segmentation algorithm are included in the package. We implemented the X-Y cut algorithm by Nagy and the Docstrum algorithm by O'Gorman. Professor Kise provided us his implementation of Voronoi-based algorithm. We also implemented the Textline-based benchmark algorithm for error counting and scoring and the Simplex-based optimization algorithm for automatically select a set of optimal page segmentation algorithm parameter values. The version of the DAFS library used in the PSET package is Release 1.0alpha0. The version of the TIFF library is v3.0. These libraries are included in the PSET package. They can be downloaded from "ftp://ftp.cfar.umd.edu/pub/contrib/sources/illuminator/illuminator-vers". The directory structure of the PSET package is described as follows: - \bin keeps all executables - \dafs keeps DAFS library and TIFF library files - \doc keeps all documentations of the PSET package - \experiments keeps experiments related files - \include keeps the PSET header file - \lib keeps the PSET library - \samples keeps several sample application of the PSET API - \src keeps all PSET source code - \tools keeps five tools generated using the PSET API Please read the README file under each directory for detailed package information, also go to doc directory for related documentations. and publications. Song Mao kanungo@cfar.umd.edu