Overview

In this first international document image binarization contest, the general objective is to record recent advances in document image binarization using established evaluation performance measures. We plan to create a benchmarking dataset that is representative of the potential problems which are challenging in the binarization process and use a common evaluation platform in order to test and compare the submitted algorithms for document mage binarization. The DIBCO 2009 dataset will contain images that range from gray scale to color, from machine printed to handwritten, and finally, from real to synthetic. Sample image documents along with the corresponding ground truth will be given to all participants right after their registration.