How to segment handwritten text forms into separate sections and words?

I want to segment image of various forms which are handwritten, into separate sections and words. I don't want character recognition; just given an image I want to segment it into a开发者_如何转开发ll the different handwritten words it contains and if it contains multiple sections (separated by lines), then a separation of those. What are the specific vision algorithms for this? Can I use any open code libraries for this. Language of the code is not a concern.

You will need a classifier combined with a parts based model. See this publication- A Shared Parts Model for Document Image Recognition - for more details.
Bear in mind, this is an active field of research.