Script to crop text area from scanned document image

I need a python script which can extract a text area from a scanned pdf. Here is a sample of example documents to test on: [login to view URL] I need the "first defendant" text area of the table cropped out and saved as a new image file. The script must work independent of the dimensions of the document image. I provided example images, the red box is what I want cropped as a new image.-