Ishchenko O. Image segmentation methods in intelligent systems for processing scanned documents

The dissertation is devoted to the solution of the actual scientific and technical problem of the developing image segmentation methods in the intelligent systems for processing scanned documents (ISPSD). The work is based on measures to image processing time reduce in the ISPSD while high quality of segmentation is saved. The intelligent systems, models and image segmentation methods in the scanned document image processing area were analyzed. A two-level image representation model of the scanned document was developed, what help to account structural properties of homogeneous regions. It allowed to develop the time-efficient processing and segmentation quality methods for identifying areas by the statistical and geometric features in photos and graphics, and to apply the multi-level image processing. The segmentation module for ISPSD Cognitive PDF/ A has been developed and its practical test with a positive result has been carried out.