Roi Framework for Generic Handwritten Amharic Form

For a decade, researchers have been working on Amharic character recognition, mainly due to large collection of paper documents in institutions and organizations. However, despite the significant achievements registered in the Amharic optical character recognition in terms of performance there is no real world implementation of the research result on a paper document collection. One of the reasons is the difference in layout structure of paper documents. This thesis explores suitable Document Layout Structure Identification Framework of Region of Interest recognition for Generic Amharic handwritten Form. The scope of the research only considers forms, which have Amharic user data entry area. A review of previous work in document layout analysis and form layout recognition was considered. Preprocessing activity was used to select appropriate forms for the experiment. Region of Interest Recognition for Generic handwritten Amharic Form framework is developed. The framework has three components. Form layout design, Form Layout Recognition, and Database was used to store information.