Bank Check Image Binarization Based on Signal matching

In this paper, a method based on signal matching to binarize low signal-noise rate bank check image is proposed. This method can extract information from the check image interfered with both complex background and imprinted seal. With the prior knowledge, the image projection function without noise is the source signal, the projection function of image binarized by iterative threshold will match the source signal, and the threshold which projection function matches best is the optimum threshold. Experimental results showed that significant improvement in the binarization quality in comparison with other well-established algorithms