On the Modification of Binarization Algorithms to Retain Grayscale Information for Handwritten Text Recognition

摘要: The amount of digitized legacy documents has been rising over the last years due mainly to increasing number on-line digital libraries publishing this kind documents. vast majority them remain waiting be transcribed provide historians and other researchers new ways indexing, consulting querying them. However, performance accuracy state-of-the-art Handwritten Text Recognition techniques decreases dramatically when they are applied these historical This is typical paper degradation problems. Therefore, robust pre-processing an important step for helping further recognition steps. proposes take existing binarization techniques, in order retain their advantages, modify such a way that some original grayscale information preserved considered by subsequent recognizer. Results reported with publicly available ESPOSALLES database.