Fine tuning Tesseract page layout analysis

[dmi...@ciitizen.com]

Is there a way to fine tune the page layout engine that Tesseract is using in order to better segment very complex and domain specific layouts? Especially when it comes to training tesseract to properly extract tabular data.