Abstract
This work is focused on the most commonly used binarization method: Sauvola’s. It performs relatively well on classical documents. However, three main defects remain: the window parameter of Sauvola’s formula do not fit automatically to the content; it is not robust to low contrasts; it is non-invariant with respect to contrast inversion. Thus, on documents such as magazines, the content may not be retrieved correctly which is crucial for indexing purpose. In this paper, we describe how to implement an efficient multiscale implementation of Sauvola’s algorithm in order to guarantee good binarization for both small and large objects inside a single document without adjusting the window size to the content. We also describe on how to implement it in an efficient way, step by step. Pixel-based accuracy and OCR evaluations are performed on more than 120 documents. This implementation remains notably fast compared to the original algorithm. For fixed parameters, text recognition rates and binarization quality are equal or better than other methods on small and medium text and is significantly improved on large text. Thanks to the way it is implemented, it is also more robust on textured text and on image binarization. This implementation extends the robustness of Sauvola’s algorithm by making the results almost insensible to the window size whatever the object sizes. Its properties make it usable in full document analysis toolchains.