Runlength Smearing algorithm is used at the beginning of document segmentation here.After the Runlength Smearing

the document image is divided into many small blocks called connected compoments.How to classify and combine those compoments correctly is the most import thing in document segmentation.A set of algrithms including Runlength smearing

how to select the smearing parameters

the classification of the connected compoments and the combination of the text blocks etc.are presented.