Abstract
The accuracy of today's document recognition algorithmsfalls abruptly when image quality degradeseven slightly. In an effort to surmount this barrier, researchershave in recent years intensified their study ofexplicit, quantitative, parameterized models of the imagedefects that occur during printing and scanning.I review the recent literature and discuss the formthese models might take. I give a preview of a largepublic--domain database of character images, labeledwith ground--truth...

This publication has 7 references indexed in Scilit: