region extraction from a text document Image

I have a image(binary image already extracted from Input Image) in which I have to extract the paragraph for example from"Apart from my to....from me"this paragraph I want to extract and similarly "during the war to...German reparation" how should I do I have extracted the body from input Image

[praveen rai]

any suggestion anyone want to give or anything addtional should I add so that anyone can help me please tell me

[KSSV]

Try ocr...

Answer 1

[Image Analyst]

You need to scan your document with something like stdfilt() with the proper size window to identify blocks of text. It will have high values, whereas smooth areas like photos and white space will have low values, Then you can threshold the stdfilt image and use bwareaopen() to look for appropriate sized areas or use regionprops and look for rectangular shaped areas.

[Image Analyst]

It will get you the entire block of text. If you want to find paragraphs within that, it looks like you can scan down the image, getting the block sum as a function of column number. The columns that start paragraphs will have a lower sum since there are no pixels there.

Answer 2

[KSSV]

I = imread('Page_body.png'); 
results = ocr(I);
str = results.Text 

[praveen rai]

its giving error on line "str= results.Txt(saying no appropriate method)

basically I want bounding box on those two paragraph in that image

[KSSV]

Typo error...it should be

str = results.Text 

Edited the code.

[praveen rai]

sorry I typed here wrong but i copy your code and paste as it is so error is giving my matlab

[KSSV]

i have dited thw code....try again..

[praveen rai]

ya thnx but what u sent I dont want that