A Multi-stage Adaptive Binarization Scheme for Document Images

In this paper, we will present a multi-stage adaptive binarization scheme for document images taken under challenging illumination condition. The approach includes two phases. In the first phase, the document image is roughly separated into background part and text part based on a coarse adaptive method. In the second phase, based on the information obtained in the previous stage, the document image is iteratively binarized using local thresholds computed based on local information. Results are presented to illustrate the algorithmpsilas effective performance for document images.