I recently worked on extraction of text from images in perl using tesseract.

Below is the approach I followed and worked well for me:

1. Download executable of tessract-OCR. Best way to do that is using SVN link for download. Need to follow all steps in
SVN link

2. It also has dependency on Leptonica

3. Once everything goes well install 'Image::OCR::Tesseract' from CPAN. If 1 and 2 works fine, all tests should pass and the installation should be successful.

Let me know if this works for you.