Paper is devoted to the construction of the method of automatical detection of text regions on image of a document using regions merging. Proposed approach combines the simplicity of merging algorithm implementation with an opportunity of removing image elements which do not contain text information. Experimental investigations were performed, classes of practical tasks to use proposed approach in the most effective way were distinguished.
text region, graphic region, merging, image of a document, threshold limitation, Hough transform
"Detektyrovanye tekstovыkh oblastei na yzobrazhenyy dokumenta metodom slyianyia" [The detection of text regions on image of a document using merge method],
Information Processing Systems,