Search for algorithms here: http://www.visionbib.com/bibliography/contentschar.html#OCR,%20Document%20Analysis%20and%20Character%20Recognition%20Systems
And I don't know the difference between your definitions of text recognition and OCR. Do you mean like you just want to find regions in the image that have letters in them but don't care what they are? Like road sign/navigation issues, or license plate recognition, or identifying photos and text blocks in a newspaper or magazine?
