Um novo algoritmo baseado em entropia para filtragem da interferÃncia frente-verso / A New Entropy-Based Algorith for Back-to-Front Interference Filtering

AUTOR(ES)

JoÃo Marcelo Monte da Silva

DATA DE PUBLICAÇÃO

2005

RESUMO

The digitalization of documents originally printed in paper is the most efficient way one can find today to preserve their contents to future generations and to make them widely accessible including their dissemination via computer networks. The particular features of each set of documents impose different techniques to document storage and digitalization. In general, to allow future possibilities documents are digitalized in true color (16M colors), and at a high resolution (reaching today over 1,000 dots per inch). Aiming at widespreading document information through network access, documents are generally available in their monochromatic version, scanned with 200 dpi resolution and compressed in a convenient format, normally TIFF (G4). The process of decreasing the palette of documents to monochromatic is known as binarization. Whenever a document is written or printed on both sides of translucent paper, there is a back-to-front interference. The standard binarization algorithms present at commercial tools generate images where the ink the front and back is overlapped, making unreadable the image obtained. Although this problem is over a decade old, better solutions to this problem are still of interest today. In historical documents, paper aging is a complicating factor. This dissertation proposes a new algorithm based on entropy of the image histogram to binarize historical documents with back-to-front interference. The proposed algorithm is compared with its predecessors described in the literature, yielding better quality images

ASSUNTO(S)

digitalized document analyses entropia engenharia eletrica interferÃncia frente-verso e imagens monocromÃticas back-to-front interference and monochromatic images anÃlise de documentos digitalizados binarization entropy binarizaÃÃo

Documentos Relacionados

• An Entropy-Based Statistic for Genomewide Association Studies
• Nonparametric Entropy-Based Tests of Independence Between Stochastic Processes
• Entropy-based gene ranking without selection bias for the predictive classification of microarray data
• Electromagnetic interference in the band of cellular mobile telephony
• EstatÃstica de contagem de carga e efeitos de interferÃncia em sistemas mesoscÃpicos