New Adaptive Methods Applied to Printed Word Image Binarization

S. Nomura (Japan), K. Yamanaka (Brazil), and O. Katai (Japan)


– Image Processing, Adaptive Lighting, Image Enhancement, Binarization, Degraded Images.


An adaptive image processing method to improve the binarization quality of the printed word images is proposed and implemented. The method is compared with a nonlinear adaptive filter method and Otsu binarization method. The Otsu method is also applied to automatic selection of binarization threshold. The data we used were degraded images that were taken from the real world where the objects may be in movement and the condition of the illumination was not sufficient. In order to show the efficiency of the new adaptive methods, the obtained results by applying a proposed system are presented. The results were satisfactory as we expected. Those results can be applied successfully helping an automatic process of characters segmentation and features extraction, contributing to increase the rate of word recognition.

Important Links:

Go Back