Return to Article Details Automatic analysis of old documents: taking advantage of an incomplete, heterogeneous and noisy corpus Download Download PDF