On Tue, 21 Jun 2005 Jens Gulden <m...@jensgulden.de> wrote: > Janusz S. Bien schrieb: > > The original scans converted to DjVu > > DjVu looks interesting (http://www.djvuzone.org/). > Thanks for that hint, I didn't know it before. > > > Although the processing improved the images in some respects, in > > general the result seems to me less readable then the > > original. > > The dirty artefacts around letters are strange. Do you think they > originate from unpaper's processing?
My wild guess is that unpaper was confused and treated parts of background as belonging to letters. > I placed an unprocessed page from your djvu-file to > > http://user.cs.tu-berlin.de/~jgulden/unpaper/archive/20050621JanuszBien/03.pbm > > and processed it with > > unpaper -vv --layout single --mask-scan-threshold 0.4 > --black-threshold 0.3 --border 0,100,0,0 03.pbm 03up.pbm > > The resulting file is > > http://user.cs.tu-berlin.de/~jgulden/unpaper/archive/20050621JanuszBien/03up.pbm Thanks. > > Do you get the same result on your system from the same unpaper run? I will resume my experiments after the return from holidays, i.e. in the end of July. Thanks once again for your help. Best regards Janusz -- , dr hab. Janusz S. Bien, prof. UW - Uniwersytet Warszawski (Katedra Lingwistyki Formalnej) Prof. Janusz S. Bien - Warsaw Uniwersity (Chair of Formal Linguistics) jsb...@mimuw.edu.pl, jsb...@uw.edu.pl, http://www.mimuw.edu.pl/~jsbien/, http://www.klf.uw.edu.pl
