Design algorithm for text and figures separating from printed image document

Abstract EN

The aim of this research is to create algorithm that has ability to separate figures and texts which are exist in document.

The foundation of this algorithm, will deal with document as an image, then separate all figures which have many colors differ from the text color.

But there is an effective problem that we might be false, there is some gray-levels values, which be near from gray-levels that exist inside the error text, whereas the error doesn't contain on black color absolutely.

Algorithm will make table consist of two fields ; the first, the colors which exist in document, and the second number repetition for each color in document, algorithm will deal in intelligence with separate white, black colors and color (gray-levels) that mixed with texts in the writing, after that insulate all gray-levels which exist in the figures.

At the final the outputs of algorithm are two documents : The first, contains only writing (texts) and the second contains figures without making some distortion on writing documents.