Improvement of Binarization Method Using a Water Flow Model for Document Images with Complex Backgrounds

摘要

Binarization algorithm using a water flow model has been presented, in which a document image is efficiently separated into two regions, characters and backgrounds, due to the property of locally adaptive thresholding. However, this method has not decided when to stop the iterative process and required long processing time. Moreover, characters on poor contrast backgrounds often fail to be separated successfully. In the current paper, an improved approach is proposed to overcome above shortcomings of the existing method, by introducing a hierarchical thresholding technique as well as extracting the regions of interest (ROIs) for speed-up and an automatic stopping criterion.