John K.'s profile - activity

I am trying to segment warped text lines found in photos taken from a page of a book (from the camera of a smartphone). This is meant as a preprocessing step before running OCR on them.

I have found two relevant papers on this:

https://sites.google.com/a/iupr.com/b...

https://www.researchgate.net/profile/...

The papers are too hard for me to implement, can anyone help? Similar methods to achieve the same result would also be great.

Example input no.1:

Example input no.2:

I want to select the staff lines as individual blobs in either of these images:

http://i.stack.imgur.com/7LJ25.png

http://i.stack.imgur.com/V95Le.png

The staff lines have similar darker gray color than the rest of the image.

Is this possible?

Edit: The regions that I want to select are shown in this image (roughly, I selected them manually): http://i.imgur.com/4bpZLNJ.png It's ok if other blobs are getting selected too (I can filter them by their area afterwards).